activity based costing and management applied in a hybrid decision support system for order...

pp

thert, M

ws bepreynaoratis cy ofDStte

2011 Elsevier B.V. All rights reserved.

play aenvironassist

morepply Chdiffere

Decision Support Systems 52 (2011) 142156

Contents lists available at ScienceDirect

Decision Supp

.eThe rst two modeling approaches emphasize the capacityavailability in order to decide whether to accept or reject an order,whereas the PTP approach considers the opportunity cost of acceptingor rejecting an order as a main decision evaluation factor. In fact, PTPexamine decisions based on the possibility of assigning the availablecapacity by not accepting an order today to another order with higherprot margin which has been predicted to take place in the future.

Regardless of themodeling approach used,management requires acomplementary tool that can assist them to analyze the impact of thedecision implemented on the business status changes. For instance,

maintenance, security, safety) which represents a challenge to assignthem to the cost objects.

The Traditional Cost Accounting (TCA) allocates, as opposed toassigning, MOH costs either by using a plant-wide rate or depart-mental rates; either case may distort the nal production costamount. Especially in a case where there is a highly customized andlow volume production process. Unrealistic cost estimation, may leadto mispricing and compromise the rm's growth and protability.

Activity-BasedCosting andManagement (ABC/M) is a relatively newcost accounting and management approach that enhances the level ofwith respect to ATP and CTP, managemeavailability of their production resources afHowever, PTP management needs to monitand prot changes after taking any dec

Corresponding author. Tel.: +1 514 848 2424x7222E-mail addresses: [email protected] (A.H.

[email protected] (A.A. Bulgak), jjsegovia@jmsb

0167-9236/$ see front matter 2011 Elsevier B.V. Adoi:10.1016/j.dss.2011.06.003nt theoretical modelingpable-To-Promise (CTP),

(MOH). The rst two are categorized as direct costs, which aretraceable to a specic cost object (e.g. service, product, order). Thelatter represents a mixture of both direct and indirect costs (e.g.approaches: Available-To-Promise (ATP), Caand Protable-To Promise (PTP).Cost control

1. Introduction

Decision Support Systems (DSSs)rigorous global competition businesstime and reliable information, DSSmaking in rendering the companyresponsive, and agile. In the area of SuDSSs can be formulated through threecrucial role in today'sment. By providing on-management decision

protable, leaner, moreain Order Management,

traditional optimization-based models cannot fulll this requirement.They are only able to provide relevant information based on thebusiness static status.

Moreover, the PTP model needs to be developed based on anaccounting approach which can accurately estimate the value ofconsumed resources. Generally, a production process requires threeinputs: Direct Labor, Direct Material, and Manufacturing Overheadnt needs to know theter fullling each order.or and control the costsision dynamically. The

understanding aABC/M is an accoMOH costs to theeliminate MOH ccosts (e.g. facility

The importancomprehensivelyAs real-life bu

; fax: +1 514 848 3175.Khataie),.concordia.ca (J.J. Segovia).

ll rights reserved.Supply chain managementMixed-Integer ProgrammingSystem Dynamics

analysis. This will lead to beActivity-Based Costing and Management afor order management

Amir H. Khataie a,, Akif A. Bulgak a, Juan J. Segovia b

a Department of Mechanical and Industrial Engineering, Concordia University, 1515 St. Cab John Molson School of Business, Concordia University, Montreal, Canada, 1450 Guy Stree

a b s t r a c ta r t i c l e i n f o

Article history:Received 8 July 2010Received in revised form 25 February 2011Accepted 10 June 2011Available online 25 June 2011

Keywords:Activity-Based Costing and ManagementDecision Support System

This article introduces a neManagement. This model ibusiness environment. Thewhich integrates System DManagement model incorptwo models, MIP and SD. Thcan evaluate the protabilitexisting optimization-based

j ourna l homepage: wwwlied in a hybrid Decision Support System

ine St. West, Montral (Qubec), Canada, H3G 1M8 (ofce EV 13105)ontral (Qubec), Canada, H3G 1M8 (ofce MB 14229)

Cost Management and Decision Support System (DSS) applicable to Ordertter t and compatible with today's competitive, and constantly changing,sented Protable-To-Promise (PTP) approach is a novel modeling approachmics (SD) simulation with Mixed-Integer Programming (MIP). This Orderes Activity-Based Costing and Management (ABC/M) as a link to merge theombination is introduced as a hybrid Decision Support System. Such a systemeach Order Fulllment policy and generate valuable cost information. UnlikeS models, the presented hybrid modeling approach can perform on-time costr business decisions based on the updated information.

ort Systems

l sev ie r.com/ locate /dssbout business operation costs; especially MOH costs.unting approach which assigns, instead of allocating,activities. Although the application of ABC/M does notosts allocation, it can reduce it to some facility-levelutility costs, facility managing costs).ce of hybrid Supply Chain DSSs has been shownin a recent study presented byMartinez-Olvera [24].

siness environments have become really complex,

decision, and monitors their business competitiveness factors dy-namically. SD is a simulation approach that was developed in the mid50s by J. Forrester from Massachusetts Institute of Technology (MIT)to understand the dynamic behavior and status alternation ofcomplex systems over a certain period of time with learning ability.

Lately, SDhas beenappliedonnumerousdiverse areas of researchbyupcoming the advancedgenerationof SDsimulation software.However,the surveypresentedbyBraines andHarrison [3] showed the limitations

143A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156supply chains members have been forced to use hybrid businessmodels (that is, the integration of features of two different businessmodels). The other three studies which have paid attention to thissubject are: [19], [20], and [26]. Martinez-Olvera [24] also discussedthe optimization-based simulation models as a potential future work.

This article explains how ABC/M can be utilized as a powerfulapproach to develop a hybrid Mixed-Integer Programming (MIP) andSystem Dynamics (SD) Decision Support System for Order Manage-ment problems. It also introduces a new approach in integratingABC/M information with SD simulation modeling technique, whichresults in a more reliable and precise cost monitoring tool.

The rest of this article is organized as follows: Section 2 provides abrief literature review, Section 3 elaborates the discussed generalOrder Management problem, Section 4 incorporates the ABC/M-basedMixed-Integer programming (MIP) decision support model, andSection 5 provides an illustrative numerical example for thedeveloped MIP model. The System Dynamics (SD) cost monitoringmodel and the hybrid DSS all are explained in Section 6. In Section 7the relevant conclusion and future work are explained. Finally, the SDmodel variables' information is given in Appendix.

2. Background

ABC/M is a two-stage process, (1) associating cost to resource(activity), and (2) selecting an appropriate activity measurement(activity cost driver), [6]. Kee [14] named the two steps as; (1) breakingoverheadcosts into different cost pools and (2) assigningoverheadcoststhrough different activity cost drivers to products or orders. As a result, amore accurate overhead costs assignment is achieved.

ABC/M supporters highlight two principal objectives, [11] and[27]: (1) to provide detailed information about the costs andconsumption of activities in a specic process and (2) to provideaccurate information for managers to improve decisions. This has alsobeen corroborated by Gosselin [7] regarding a pilot and full ABC/Mimplementation studies. However, the use of ABC/M has been limitedto a cost accounting approach, rather than as a managerial technique(Gosselin [7]; Kaplan and Anderson [13]; Gosseling [8]).

ABC/M advantages, and constructive effects, on a rm's perfor-mance have been determined through numerous studies anddissertations. Kennedy and Aeck-Graves [15], Ittner et al. [12], andCagwin and Bouwman [5] attested ABC/M as a preferable accountingapproach compared to the TCA systems. Some studies such as;Novievi and Anti [25] and Cagwin and Barker [4] showed evidenceof a positive impact of ABC/M on lean manufacturing components likeJust-In-Time (JIT) and Total Quality Management (TQM).

The preeminence of ABC/M in providing detailed cost informationrepresents a potential powerful approach for developing PTP SupplyChain Decision Support Systems. Malik and Sullivan [22] developed anABC/M-based Mixed-Integer Programming (MIP) decision supportmodel for product mix problems. Kee [14] integrated some aspects ofthe Theory of Constraints (TOC) in ABC/M-based MIP modeling forthe product mix problem and named it Expanded ABC/M model.The model identies the rm's optimal product mix by evaluatingsimultaneously the resources and product cost, the productionresources availability, and the business marketing opportunities.

In Supply Chain Order Management, [17] and [18] presented a PTPMIP model for accepting or rejecting orders by implementing ABC/Mhomogeneous cost pools' structure originally introduced by Cooper andKaplan [6]. The purpose of the model was to gain insight into howsignicant OrderManagement decisions are inmaximizing protabilitywhen the rm has insufcient production resources to satisfy all thedemand. Khataie et al. [16] added the possibility of pursuing two maindifferent goals simultaneously, reducing the residual capacity andincreasing the protability to the previous models.

A powerful PTP Order Management tool assists management to

monitor, analyze, and foresee the consequences and outcomes of eachof SD modeling in the manufacturing sector from the business and/oroperational perspective. This represented a diversication from itsoriginal purpose, which was to serve as a decision support tool formanufacturingprocesses at the operational level. Instead, and accordingto the survey, SD have been broadly used in the modeling of resourcemanagement at national and global level decision support processes,and in the service sector at operational levels.

A limited number of studies using SD approach for nancialdecisions have been reported. Abdel Hamid and Madnick [1] used SDfor software development cost estimation. They implemented the SDsimulation technique to see the effects of multi-variable changes inthe model. Marquez and Blanchar [23] applied SD to supportinvestment decisions in high-technology business. Macedo et al.[21] developed a real-time cost monitoring model for the reengineer-ing process of a gelatinous substance at the microbiology laboratoryby integrating the ABC/M and SD. However, the developed model wasacting as a real-time cost calculator rather than a System Dynamicsmodel. There were no positive or negative learning loops in theproposed model.

3. General illustrative problem

A Flexible Manufacturing System is selected as a pilot productionfacility in a simplied three-echelon Supply Chain including supplier,producer, and customers. The system can set up two different productionprocesses or models: Basic and Deluxe. Direct material is similar forboth models and there is no restriction for direct material supply. Themanufacturing process, Fig. 1, starts by injecting the common directmaterial into the system. Second, the FMSalternatesbetween two typesofsetup based on the assigned production plan. Lastly, the manufacturingproducts are stored for shipping to the customers.

The rm's management follows a pull-production strategy,therefore it develops the aggregate production plan (AP) based onthe received orders per month. Not all the orders can be fullledcompletely due to machining hour capacity per period; as a result, therm's management has to choose the fulllment rate of each order.The Order Management policy is fullling completely, or partially, orrejecting the orders according to the production system availabilityand order's protability factors.

The problem and parameters have been extracted from amanagerial accounting educational business case study known asWillow Company from [10]. The production costs have been splitinto two groups; prime costs (which include Direct Materials andDirect Labor) and overhead costs. The latter is divided into vehomogeneous cost pools with a particular activity cost driver for eachone. The case study assumes that the overhead unit-level costs arecompletely traceable and are included in the prime costs.

There are two different batch-level cost pools introduced in thecase study; material handling and setups. Their respective cost drivers

FMSRM

PBasic

PDeluxeFig. 1. Manufacturing process ow.

are: number of moves and number of setups. The case also presentstwo order-level cost pools: administrative cost pool with activity costdriver of number of orders and engineering supports cost pool withactivity cost driver of maintenance hours. The last pool is facility-levelwhich has a unit-level activity cost driver, machining hours, based onthe hint given in the case study. Fig. 2 shows the activity-based costow down diagram for Willow Company.

4. Order Management MIP decision support model

The objective of the Order Management DSS is to nd the mostprotable and optimal combination of the fullling ratio of thereceived orders. This should be accomplished by taking into accountthe orders' protability, the production resources productivity andavailability. In ABC/M the overhead cost charged to each product isdetermined by identifying the actual consumption of the activity byeach product, and thenmultiplying it by its respective unit cost driver.

For the purpose of modeling the rst part of DSS, we implementedthe technique that was introduced by [16], solving the OrderManagement problem by integrating the ABC/M and MIP optimiza-tion techniques. The integration can be done by applying Cooperand Kaplan [6] homogenous cost pooling structure. According totheir structure, the overhead costs can be assigned to four specichomogenous cost pools; unit, batch, product, and facility levelactivities' cost pool.

The unit-level activities (machine time, direct material, direct

cost in a production process in a more precise and reliable manner ascompared to the TCA approach.

The developedmathematical OrderManagement Decision Supportmodel pursues two goals: maximizing the prot and minimizing theresidual capacity. These two goals are incorporated into the modelby applying Weighted Goal Programming (WGP) techniques. Thegeneral objective function is maximizing the prot (sales revenueproduction resources costsholding costs) and minimizing theresidual capacity simultaneously, subject to different constraints likeproduction resources constraints, order commitment constraints,management discretionary constrains, and inventory constraints.The management discretionary factor is also added to the model byusing a constraint that fullls a certain number of orders completed.

The generic model is developed based on the following assump-tions: processing times are deterministic, each product is manufac-tured in equal-size batches under a pull system, and each orderconsists of just one type of product. There is a possibility to satisfy theorders partially, completely, or even reject them.

In this study, the general homogenous costs' pooling structure,which has been used in previous studies, is replaced by a more detailedhomogeneous costs' pooling structure. Instead of grouping all the batch-level activities in one cost pool, we are dealingwith two different batch-level overhead cost pools; similarly for order-level activities. In addition,the facility-level overhead costs' pool is also considered in this casestudy. As mentioned before, for the Willow Company case study, theunit-level overhead activities' resources and costs are integrated into

Ind

No

F

Or

144 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156labor, etc) cost pool includes the overhead costs that vary directlywith the number of units produced. The batch-level activities(planning and tactical management, material handling, setup, etc.)cost pool which are the overhead costs invoked whenever a batch isprocessed. The product-level activities (process engineering,manufacturing equipments maintenance, product design, etc.) costpool which are overhead costs incurred whenever a particularproduct is produced. In this study, the product-level is replaced byorder-level which is overhead costs incurred whenever a particularorder is processed. The facility sustaining activities' cost pool includesoverhead costs such as rent, utilities, and facility management.Integrating the ABC/M approach into the Order Management DSSshows and claries the role, importance, and source of each overhead

Direct Costs

Mat

eria

lH

andl

ing

Setti

ng U

pEq

uipm

ents

Purc

hasin

gM

ater

ials

No. Of Moves No. Of Setups

Define the Order

Act

ivity

Cos

t Dri

ver

Cos

t Poo

l

Prim

e C

osts

+H

oldi

ng C

ost

+ D

irec

t Ove

rhea

d C

osts

Batch-Level 2Batch-Level 1Fig. 2. Activity-based cost ow dithe prime costs. The applied notations are as follows:

i product indext period of time indexo order indexj activities at unit-level indexk activities at batch-level indexl activities at order-level indexr activity at facility-level indexdl+ amount of over capacity production (capacity surplus

variable)dl- amount of under capacity production (capacity slack

variable)

irect Costs

Payi

ng S

uppl

ier

Rec

eivi

ngG

oods

Engi

neer

ing

Supp

ort

Mai

ntai

ning

Equi

pmen

t

Prov

idin

g Sp

ace

. Of Orders MaintenancesHoursMachining

Hours

ulfillment Cost

Order-Level 2der-Level 1 Facility-Levelagram for Willow Company.

145A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156pri prime cost of product iak batch-level k pool rateyl order-level l pool ratecr facility-level r pool ratehi holding cost of product i per periodqijr consumption rate of performing activity r at facility-level

related to activity j at unit-level for product iuijk consumption rate of performing activity k at batch-level

related to activity j at unit-level for product ifil consumption rate of performing activity l for product ibij batch size of product i at activity jpi sales price of product im1 cost of stretching the production capacitym2 cost of not using whole capacitym3 preference coefcient of maximizing protm4 preference coefcient of minimizing residual capacityO number of orders which should be fullled completelyDiot demand quantity of product i in order o due in period tQjt total available time to perform activity j in period tUkt total available time to perform activity k in period tFl total available time to perform activity lIit inventory level of product i in period tPit amount of product i produced in period iSiot quantity of product i from order o sold in period tBijt number of batches of product i produced in machine j in

period tYiot the proportion of accepted order o from product i in period tYBiot the binary form of Yiot

In order to develop the multi-objective function we used WGPtechnique. Eq. (1) represents the objective function, which consists oftwo parts that pursue two different goals of the Decision Supportmodel. The mi represents each goal's signicance or managementpreference coefcient for each goal. The rst part calculates the prot,which is the revenue (multiplication of sales by the product price)minus the production process costs. The production process costs isthe addition of the prime costs, overhead costs (batch-level, order-level, and facility-level) and product's holding costs. The second part ofthe objective function minimizes the residual capacity. According tothis term any violation from the available Order Fulllment capacityhas a certain penalty cost.

Max z = m3 toipi Siottipri

Pittijkak uijk Bijt

toilyl fil Yiottoijrcr

qijr Siotithi Iitm4 lm1dl + m2dl

1

Constraints (2) and (3) ensure that the available Order Fulllmentresource capacity at unit-level and batch-level capacity, respectively,are not exceeded. Constraints (4) allow the diversity in batch sizes toexist.

iqijy PitQ jt j; r; t 2

ijuijk BijtUkt k; t 3

Pit = bij Bijt i; j; t 4

Constraints (5) make sure that the production quantity meets thesales commitments of each order. Constraints (6) are the order-levelactivities capacity variation constraints. They ensure that we accept

the orders considering the order-level activities available capacity.The ratio of Order Fulllment is represented with the proportionfulllment decision variables (Yiot). These decision variables can takeany real number between 0 and 1, which represents the acceptanceportion of each order.

Siot = Diot Yiot i; o; t 5

iot fil Yiotdl + d

l = Fl l 6

Constraints (7) incorporate the management discretionary factorinto the model, which represents the number of orders (O) thatmanagement decides to fulll completely. Constraints (8) ensure thatif a specic order is selected to be fullled completely then therelevant fulllment ratio (Yiot) is equal to 100%. Constraints (9) makesure that the feasible area only includes the orders that exist.Accordingly, if the demand of a certain order from certain productat certain period (Diot) is equal to 0 then Yiot of that order must beequal to zero.

iotYBiotO 7

YiotYBiot i; o; t 8

DiotYiot i; o; t 9

Constraints (10) and (11) determine the inventory level for eachproduct type at the end of each period of time.

Ii0 = 0 i 10

Ii t1 + PitIit =oSiot i; t1 11

Finally, constraints (12) to (14) are the non-negativity constraints,and constraints (15) are the binary constraints.

Pit0 i; t 12

Bijt0 i; j; t 13

0Yiot1 i; o; t 14

YBiot = 0 or 1 i; o; t 15

The model has an illustrative emphasis on the effect of overheadcosts in the decision making process by replacing the homogenousorder and batch-level overhead costs pool with a more detailed andactivity oriented overhead cost pools. It also integrates the facility-level activities overhead costs pool. In fact, the model elaborates onhow the signicant role of overhead costs are in the procedure ofOrder Management decision making in a better way as compared tothe previous MIP Order Management models. In the following sectionthe model is validated by using a numerical example.

5. Numerical example

In this case study, management is required to make decisionsregarding sixteen received orders over the next twelve periods of time(months). The list of the received orders and their specications areshown in Table 1. Each order consists of only one type of producteither the Deluxe or the Basic model. The objective is to nd theoptimal combination of Order Fulllment ratio, the combination thatmaximizes the prot and minimizes the residual capacity for thesystem. All the required nancial and operational parameters areextracted from the Willow Company case study.

Even though the Willow Company case study does not includeholding costs, we decided that such a variable is important in a

realistic scenario. Therefore, a holding cost has been assigned to each

product. An applicable holding cost (hi) is assumed to be equal to 5% of

facility-level costs (xed overhead costs) by relaxing the constraint offullling a certain number of orders completely.

The related optimal solution is by fullling ve orders completelyand six orders partially. The table also shows the effect of integratingthe management discretionary factor into the model where a certainnumber of orders have to be fullled completely. Managementstrategy for reaching higher customer satisfaction may require morenumber of orders to be fullled completely. According to Table 2,demanding more than ve completely fullled orders diminishes thecompany's prot. In fact, management would sacrice the short-termprot for having a higher customer satisfaction level and long-termprot. This is the consequence of having the set of constraints (7) as abinding constrain.

The developed PTPmodel, by integrating the ABC/M information intothemathematical OrderManagementDecision Supportmodel, elaboratesmore on the role of costs and especially overhead costs in the OrderManagement process compared to the more traditional CPT and ATPmodels. However, the presentedMIPmodel solely cannot provide an on-time detailed cost analysis for the different Order Fulllment scenariosbecause of its static nature. In fact, it does not take advantage of all the

Table 1Order specications.

Order number Product type Period due Quantity

1 Basic 1 35002 Basic 2 46003 Deluxe 2 25004 Basic 2 30005 Deluxe 3 25006 Basic 5 32007 Basic 5 47008 Deluxe 5 19009 Deluxe 6 400010 Basic 8 450011 Deluxe 9 300012 Basic 10 350013 Basic 11 300014 Basic 12 500015 Deluxe 12 270016 Deluxe 12 5000

146 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156the ABC/M-based unit manufacturing cost as indicated in reference[10]. Therefore, the amounts used in the MIP model are 4.22 and 9.06for the Basic and the Deluxe models, respectively, in dollar/unit/month. A further elaboration of holding costs requires incorporatinginventory-related activities into the model structure.

The manufacturing facility can produce the Basic and the Deluxemodels in batch sizes of 100 and 170 units, respectively, based on thegiven projected production amount and the number of setups for eachproduct. The resource annual capacity at different level has beendetermined based on the cumulative forecasted annual resourceconsumption given in the case study for producing the projectedamount for each product. The preference coefcients (m1 to m4) areequal to 1, 0.5, 1, and 2, respectively given from [16]. According to thiscombination the goal of minimizing the residual capacity has highersignicance compared to the prot maximization goal.

The model is coded with the optimization software Lingo Version10 and applied to the different scenarios, different desirable numberof orders (O) which should be satised completely, on a 3.00 GHzPentium-4 processor with 1 GB of RAM. The average computationaltime for the presented model is less than 5 s.

As discussed before, the goal is to nd the optimal combination offullling ratio of the received orders by taking into account theproduction resources capacity availability and protability factor foreach order. According to the results shown in Table 2, the optimalvalue is $3,694,067.00 which considers all the costs, including theTable 2Comparison table for fullling rates.

Minimum desirable amount of orders which should be satised completely

Order 5 or less fullling rate % 6 Fullling rate % 7 Fullling rat

1 86 86 862 65 3 14 14 144 - 100 1005 27 27 66 72 72 727 100 100 1008 45 45 1009 17 17 1710 100 100 10011 68 68 6812 100 100 10013 100 100 10014 100 100 10015 16 20 20 20Prot $ 3,694,067.00 3,693,925.00 3,681,280.00information generated by applying ABC/Mcost structure. Therefore, thereis a need for a complementary decision support model.

The Decision Support System presented in this paper can be usedas a cost monitoring and cost analysis tool with the ability to evaluateand foresee the effects of each taken decision on the system status.The next steps explain how the output of the MIP model can be usedas an input of the System Dynamics-based cost analysis model andhow ABC/M is used as a common approach to link these two modelsand introduces them as a hybrid DSS.

6. System Dynamics decision support model

The main advantage of System Dynamics as a continuoussimulation approach versus a discrete event simulation approach isits ability to continuously update the system variables after eachdecision is taken (simulation run). Within this optic the nextsimulation run will be based on the updated variables from theprevious run. SD has been applied in a wide range of areas; corporateplanning and policy design, biological and medical modeling, energyand environment, software engineering, and supply chain manage-ment, according to the study of Angerhofer and Angelides [2].

Gregoriades and Karakostas [9] presented a framework to showhow SD can integrate with business objects and act as a powerfulsimulation approach to help organizations in pursuing their goals andmonitoring their processes. We believe the ability of SD in on-time

e % 8 Fullling rate % 9 Fullling rate % 10 Fullling rate %

60 31 No feasible solution

14 8100 10027 100 100100 10030 100

100 100100 100100 100100 100100 100

20 203,650,092.00 3,614,196.00

evaluation of the system status can be used in cost monitoring processas well. The remaining of this study is focused on presenting a novelcost monitoring model for Order Management problems and thecombinationwith the developedMIPmodel. The combination of thesetwomodels creates the hybrid Decision Support System. Fig. 3 dashedlines show the ow of information.

The SD model is developed based on the earlier variables asdened in the MIP model and with respect to the similar appliedABC/M structure. This represents ABC/M as a common approachbetween the MIP and SD decision support models. These twomodels are linked through a spreadsheet generated by the MIPmodel. The main objective of the SD model is to estimate real-timeselling price and adjust pool rates in each month, based on theprevious months' Order Fulllment policy received from the MIPmodel.

Basically, management decides about the scenario of the OrderFulllment, number of orders that should be fullled completely, andthe MIP model provides the optimal solution for the desirablescenario; this includes the Order Fulllment rates for each order, ascan be observed in Table 2. In the next step, the output of MIPmodel isused as the input for SD model in order to have the possibility of on-timemonitoring of related costs and to dene the minimum products'selling price in each period regarding the previous decisions anddesirable markup.

The SD model structure contains level variables, rate variables,and auxiliary variables which are all related to each other within

auxiliary variables to avoid the model to be crowded and unclear.The SD model is developed in Vensim software with the similarlegends discussed in [28] and it includes six different parts.

6.1. First part

The rst part of the SD model, Fig. 4, calculates the totalholding cost separately for each product by getting the exact levelof inventory for each type of product in each period and therelated holding fractional ratio for each case. All the initial poolrates and the other related constants (e.g. batch sizes, markups,holding cost) are similar to the MIP model. The Order Fulllmentrate, production rates and shipping rates are extracted from theoutput of MIP part of the hybrid model. As mentioned before,the SD part can read the MIP model output from the spreadsheetgenerated by LINGO.

6.2. Second to sixth part

All the ve learning loops that determine and adjust the poolrates are shown in Figs. 5 to 9. The pool rates' adjustment loopshelp the model to dene the selling price based on the actualOrder Fulllment costs. For the batch-level-1 (material han-dling), Fig. 5, the loops are batch-level-1 cost rate for the Deluxemodel total bacth-level-1 cost for Deluxe model total batch-level-1 costbatch-level-1 pool rate and the same loop for the

lic

eci

mat

147A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156and without ows. Level or stock variables are represented byrectangles and they show the level of discussed unit (e.g. productsor money) in the system at different periods of time. Thecombination of level variables normally denes the status of thesystem at different times. Rate variables control the pace ofchange in a specic level variable and are represented by valves inthe model; in fact, they determine the ow. The auxiliaryvariables, which are shown with clear boxes, simplify the modeland make it easier to understand. The clouds play the role ofsource and sink nodes for the in and out ows. This means the owcomes from or goes to outside boundaries of the model. Thevariables within single left and right-pointing angle quotes areshadow variables. They are substitute of any level, rate, or

Opt

imiz

atio

n M

odel

Fulfillmentrates

Fulfillment Po

Hybrid MIP-SD D

OrdersSpecification

Syste

m In

put

ABC/M inforFig. 3. Hybrid Order ManagemenBasic model. These two loops adjust the batch-level-1 pool ratein each period based on the material handling activity resourceconsumption in the previous periods. The batch level-1 pool rateis also related to the total consumption of batch-level-1 costpool activity driver. This level variable is estimated via batchlevel-1 activity cost driver consumption ratio for the Basic andDeluxe models, which eventually depends on the products'batch size, production rate, and batch-lelvel-1activity consump-tion ratio.

Fig. 6 shows the relations for the batch-level-2 (Setup). The poolrate adjustment loops are similar to the batch-level-1 loops. The totalconsumption of bacth-level-2 cost pool activity driver is estimated ineach period of time via the product's production rate, batch size, and

System D

ynamics

Model

y Spreadsheet

sion Support System

Fulfillmentrates

PricingStrategy

System O

utput

ion systemt Decision Support System.

batch-level-2 activity driver consumption ratio, which is similar to theprevious cost pool.

The relations for order-level-1, which is a homogenous cost pool,contain their different MOH costs; procurement material, paying

supplier, and receiving goods; it is presented in Fig. 7. In this case thepool rate adjustment loops are order-level-1 cost rate for Deluxemodel total order-level-1 cost for Deluxe model total cost of order-level-1cost poolcost-level-1 pool rate and the same loop for the Basic

Inventory Level ofDeluxe Model

Inventory Level ofBasic ModelBasic Model

Production Rate

Holdeing Cost Rate ofDeluxe Model

holding cost fractionalratio for Basic model

holding cost fractionalratio for Deluxe model

Total Sales ofDeluxe ModelDeluxe Model

Shipping Rate

Total Sales ofBasic Model

Basic ModelShipping Rate

Holding Cost Rate ofBasic Model

Deluxe model holdingcost coefficient

Basic model holdingcost coefficient

Total Holding Costof Deluxe Model

Total Holding Costof Basic Model

Total Production ofDeluxe Model Deluxe Model

Production Rate

Total Production ofBasic Model

Fig. 4. Inventory level and holding cost approximation.

h sxe

148 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156batcDeluTotal Batch-Level-1(Material Handling)

Cost for Deluxe Model

Total Batch-Level-1(Material Handling)

Cost for Basic Model

Batch-Level-l Cost (MaterialHandling) Rate for Deluxe

Model

Batch-Level-1 Cost(Material Handling) Rate for

Basic Model

batch-level-1 (materialhandling) pool rate

batch(num

r

to(ma

batch size ofBasic model

batch-level-1consumption

of B

Fig. 5. Batch-level-1 pool rize ofmodel

-level-1 activity cost driverber of moves) consumptionatio by Deluxe model

batch-level-1 activity cost driver(number of moves) consumption

ratio by Basic model

tal batch-level-1terial handling) cost

Total Consumption Batch-Level-1Cost Pool Activity Driver (Number

of Moves)

Total Consumption Ratio ofBatch-Level-1 Activity CostDriver (Number of Moves)

batch-level-1 activity cost driverconsumption ratio by each batch

of Deluxe model

activity cost driver ratio by each batchasic model

ates adjustment loops.

Total Batch-Level-2(Equipments Setup)

Cost for Deluxe Model

Total Batch-Level-2(Equipments Setup)

Cost for Basic Model

Batch-Level-2 Cost(Equipments Setup) Rate for

Deluxe Model

Batch-Level-2 Cost(Equipments Setup) Rate for

Basic Model

batch-level-2(equipments setup) pool

rate

batch-level-2 activity cost driver(number of setups) consumption

ratio by Deluxe model

batch-level-2 activity cost driver(number of setups) consumption


total batch-level-2(equipments setup) cost

Total Consumption Batch-Level-2Cost Pool Activity Driver (Number

of Setups)

Total Consumption Ratio ofBatch-Level-2 Activity CostDriver (Number of Setups)


of Deluxe model


of Basic model

Fig. 6. Batche-level-2 pool rates adjustment loops.

Total Order-Level-1Cost for Deluxe Model

Total Order-Level-1Cost for Basic Model

Order-Level-1 CostRate for Deluxe Model

Order-Level-1 CostRate for Basic Model

order-level-1pool rate

order-level-1 activity cost driver(number of orders) consumption


order-level-1 activity cost driver(number of orders) consumption


total cost oforder-level-1 cost pool

Total Consumption Order-Level-1Cost Pool Activity Driver (Number

of Orders)

Total Consumption Ratio ofOrder-Level-1 Activity CostDriver (Number of Orderss)

Deluxe model orderfulfillment rates

Basic model orderfulfillment rates

order-level-1 activity cost driverconsumption ratio by each order

of Deluxe model


of Basic model

Fig. 7. Order-level-1 pool rates adjustment loops.

149A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156

Total Order-Level-2Cost for Deluxe Model

Total Order-Level-2Cost for Basic Model

Order-Level-2 CostRate for Deluxe Model

Order-Level-2 CostRate for Basic Model

order-level-2 activity cost driver(maintenances hours) consumption


order-level-2 activity cost driver(maintenances hours) consumption


total cost oforder-level-2 cost pool

Total Consumption Order-Level-2Cost Pool Activity Driver

(Maintenances Hours)

Total Consumption Ratio ofOrder-Level-2 Activity CostDriver (Maintenances Hours)

order-level-2pool rate


of Deluxe model


of Basic model

Fig. 8. Order-level-2 pool rates adjustment loops.

150 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156model. In the order-level-1 cost pool, the MOH costs are estimated viaorder fulllment rates for the Deluxe and Basic models and order-level-1 activity cost driver consumption ratio.

For the order-level-2 homogeneous cost pool, engineering andmaintenance, the relations between variables are presented in Fig. 8.The relevant adjustment loops are designed similar to the order-level-1MOH costs for the Basic and Deluxe models. The total consumption ofTotal Facility-Level(Providing Space) Cost

for Deluxe Model

Total Facility-Level(Providing Space) Cost

for Basic Model

Facility-Level (ProvidingSpace) Cost Rate for Deluxe

Model

Facility-Level (ProvidingSpace) Cost Rate for Basic

Model

facility-level (providingspace) pool rate

facility(mach

ra

total co(prov

Fig. 9. Facility-level pool raorder-level-2 cost pool activity driver is estimated through orderfulllment rates for each product model and the order-level-2 activitycost driver consumption ratio.

Fig. 9 presents the relations between variables involved estimatingthe facility-level MOH cost. The pool rate adjustment loops are similarto the previous cost pools. We are considering the facility-levelactivity cost driver consumption ratio and the orders' shipping rates in-level activity cost driverining hours) consumptiontio by Deluxe model

facility-level activity cost driver(machining hours) consumption


st of facility-leveliding space) cost

pool

Total Consumption Facility-LevelCost Pool Activity Driver

(Machining Hours)

Total Consumption Ratio ofFacility-Level Activity CostDriver (Machining Hours)

facility-level activity cost driverconsumption ratio by each unit

of Deluxe model

facility-level activity cost driverconsumption ratio by each unit

of Basic model

tes adjustment loops.

Total Prime Cost ofDeluxe Model

Total Prime Cost of

Prime Cost Rate ofDeluxe Model

fractional prime costratio for Deluxe model

ofred

rderasi

total overhead cost forDeluxe model

Deluxe modelselling price

Basic modelselling price

Deluxe modelmarkup

Basic modelmarkup

Deluxe model cost of goodsmanufactured per unit

t of per

Deluxe model cost ofgoods manufactured

d pr

151A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156Basic ModelPrime Cost Rate of

Basic Model

fractional prime costratio for Basic model

Basic model costgoods manufactu

total overhead costfor Basic model

goodsunit

oduct cost estimation.by adding the specic markup for each product to the related cost ofgoods manufactured per unit for that product. The related equationsto the SD model variables are presented in alphabetical order in theAppendix.

The model can be applied to the different Order Fulllmentscenarios in order to appraise the attributes of each Order Fulllmentpolicy and evaluate the effect of the management discretionary factoron the manufacturing cost and subsequently on the selling price(selling price=(1+markup)cost of goods manufactured per unit).The variation in the selling price based on the number of orders thatshould be fullled completely is shown in Table 3. The indicated pricesin period one are the prices used by the MIP model.

The selling prices estimated by the SD model are calculatedthrough the actual manufacturing cost. The estimated prices are

ng price $ 8 Selling price $ 9 Selling price $

Deluxe Basic Deluxe Basic Deluxe

0 360.000 180.000 360.000 180.000 360.0005 347.461 178.361 347.461 177.923 347.4611 356.987 179.591 356.955 180.829 356.9995 349.049 179.933 350.215 181.999 350.8524 352.819 181.482 349.415 184.531 356.8196 361.395 185.620 349.962 188.559 366.3543 358.145 184.460 350.457 186.896 363.1996 355.685 184.658 354.435 186.817 364.4818 356.718 186.064 360.160 188.079 368.1876 360.747 185.858 368.390 187.668 374.6756 360.227 186.271 367.541 187.871 373.7051 360.302 186.403 367.384 187.828 373.3802 356.628 183.225 356.865 184.917 363.009

Table 4Total cost of goods manufactured per model.


Product 5 or less cost of goods manufactured$

6 Cost of goods manufactured$




Basic 2,343,900.00 2,344,000.00 2,256,110.00 2,359,860.00 2,289,830.00Deluxe 883,103.00 883,103.00 986,631.00 911,711.00 1,022,440.00

Table 5Total production per model.


Product 5 or less total production unit 6 Total production unit 7 Total production unit 8 Total production unit 9 Total production unit

Basic 27,000 27,000 26,000 27,000 26,000Deluxe 4930 4930 5440 4930 5440

7 T

9079

152 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156used as a reference price in implementing the Order Fulllmentpolicy, instead of the selling price used by MIP model. Thecalculated selling prices for each product per period are indeed alower limit for the products selling price. Thus, if management iswilling to achieve the desirable projected prot, it should chargethe customer no less than the calculated selling price per periodfor each type of product. According to the SD model output,

Table 6Total overhead cost per model.


Product 5 or less total overhead cost $ 6 Total overhead cost $

Basic 94,247.20 94,338.00Deluxe 71,830.60 71,830.60Table 3, fullling more orders completely would require toincrease the selling price.

The rise in the average products' selling price could be theconsequence of an increase in the total cost of goods manufactured(that could be the result of changes in the total overhead costs, totalprime costs, and/or total holding costs) and/or an increase in themanufacturing system residual capacity (this could be the result ofchanges in the production rate).Table 4 exhibits the total cost of goodsmanufactured for each model. Table 5 shows the related productionamount and Table 6 displays the total overhead costs for each producttype. Similar tables can be extracted from the model output for theother variables.

0

50

100

150

200

1 2 3 4 5 6 7 8 9 10 11 12

Dol

lar/

Ord

er

Time(Month)

Order-Level-1 Pool Rate

Fulfilling 5,6, or 7 Orders CompletelyFulfilling 8 Orders CompletelyFulfilling 9 Orders Completely

Fig. 11. Order-level acThe variations in the cost amounts are because of the selectedorder fulllment policy, which denes the inventory policy and theproduction rates. For example, in the case of 6 orders to be completelyfullled, the total production amount according to the Table 5 for theDeluxe model is 4930 units and for the cases of 7 and 8 orders to becompletely fullled are 5440 and 4930 units respectively. Therefore,the increases in both MOH and the manufacturing costs from 6 to 7 as

otal overhead cost $ 8 Total overhead cost $ 9 Total overhead cost $

,696.40 94,114.10 90,173.30,327.80 71,762.50 79,090.50well as the decreases from 7 to 8 can be justied.However, the production amount is not the only reason in cost

variations. The other reason is due to the changes in the pool rates.The model adjusts the pool rates after each run. This justies thedifference between theMOH costs for the Basicmodel from the case of5 to the case of 6, although the total production amount remains thesame. The other reason for the cost changes is due to the inventorycost which is different for each order fulllment policy.

The model also has the ability to adjust the pool rates. Fig. 11reveals the adjustment for the order-level activities pool rates indifferent Order Fulllment scenarios. The disparity between theorder-level pool rates under different fulllment scenarios is because

010203040506070

1 2 3 4 5 6 7 8 9 10 11 12

Dol

lar/

Ord

er

Time(Month)

Order-Level-2 Pool Rate


tivities pool rates.

im

-L

s Coletelete

ility

153A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156of the correlation between the Order Fulllment ratios and the order-level activities. In contrast, in Fig. 12 there is no correlation between

1.9881.99

1.9921.9941.9961.998

22.002

1 2 3 4 5

Dol

lar/

Ord

er

T

Facility

Fulfilling 5,6, or 7 OrderFulfilling 8 Orders CompFulfilling 9 Orders Comp

19.919.9219.9419.9619.98

2020.02

1 2 3 4 5 6 7 8 9 10 11 12

Dol

lar/

Ord

er

Time(Month)

Batch-Level-1 Pool Rate


Fig. 12. Batch-level and facbatch-level and facility-level activities' pool rates and the OrderFulllment scenario. Accordingly, the pool rates have not changed forthose activities at different Order Fulllment scenarios.

7. Conclusion and future work

This study introduces a novel modeling approach by integratingSystem Dynamics and MIP programming in order to develop a powerfulhybrid PTP Decision Support System for the Supply Chain OrderManagement problem. The developed DSS system assists managementin monitoring, analyzing and foreseeing the consequences and outcomesof each decision and monitors their business competitiveness factors. Inthe rst step, we developed a mathematical Decision Support modelbased on the ABC/M cost structure. In theMIPmodel amore detailed andactivity-oriented cost structure is used to enhance the model accuracy.

As a second step, the ABC/M-based SD model presented a comple-mentary tool to theMIPmodel. Thismodel identies the interconnectionsand correlations between the Order Management decision makingvariables. The SD model helps management to investigate and examinethe further consequences of executing the different Order Fulllmentdecision scenarios expansively. Themodel adjusts the pool rates based onthe actual costs, denes the on-time selling price based on the OrderManagement fulllment policy, and can also serve as a cost monitoringtoolwith thepurpose of checking the costs behavior at different levels andfor different products. ABC/M, as a common modeling approach, makestwo models work together as a hybrid Decision Support System.

The hybrid DSS output indicates that fulllingmore orders actuallydecreases the company's prot (MIP part output), and requiresadjusting the product selling price (SD part output). Depending on theproduct type and applied Order Fulllment scenario, the selling pricecould be decreased or increased compared to the initial selling priceused in the MIP model. Reducing the selling price can give moresatisfaction to the customer if the level of order fulllment remainsthe same. However, increasing the selling price may result in a lower

6 7 8 9 10 11 12e(Month)

evel Pool Rate

mpletelylyly

0200400600800

100012001400

1 2 3 4 5 6 7 8 9 10 11 12

Dol

lar/

Ord

er

Time(Month)

Batch-Level-2 Pool Rate


-level activities pool rates.or higher customer satisfaction level. This depends on the customer'sunderstanding and the value given to a better order fulllmentservice. Thus, it should be considered that the result of fullling moreorders completely actually may conict with the original intention ofincreasing customer satisfaction.

This study can be further expanded by integrating the cost ofbacklog among the decision making factors, which improves themodel accuracy level. Illustrating the role of raw material suppliersand raw material inventory into the hybrid system through theintegration of the supplier selection decision based on the suppliercosts analysis and raw material holding costs could enhance themodel legitimacy level. Another potential extension of this research isby using a similar approach to evaluate the inventory cost. Moreover,a comprehensive design of experiments can be used to analyze themodel output and price variations.

Appendix

1. Basic model cost of goods manufactured=total overhead costfor Basic model+Total Prime Cost of Basic Model+TotalHolding Cost of Basic ModelUnits: Dollar

2. Basic model cost of goods manufactured per unit=IF THEN ELSE(Total Production of Basic ModelN0, Basic model cost of goodsmanufactured/Total Production of Basic Model, 84.4)Units:Dollar/Unit

3. Basic model holding cost coefcient=0.05Units: 1/Month4. Basic model markup=1.1327Units: Dmnl5. Basic model order fulllment rates: = GET XLS DATA(FB.xls,

Sheet1, 1, B3)Units: Order/Month6. Basic Model Production Rate: = GET XLS DATA(PB.xls, Sheet1,

1, B3)Units: Unit/Month

154 A.H. Khataie et al. / Decision Support Systems 52 (2011) 1421567. Basic model selling price=(1+Basic model markup)Basicmodel cost of goods manufactured per unitUnits: Dollar/Unit

8. Basic Model Shipping Rate: =GET XLS DATA(SB.xls, Sheet1,1, B3)Units: Unit/Month

9. batch size of Basic model=1000Units: Unit/Batch10. batch size of Deluxe model=170Units: Unit/Batch11. batch-level-1 (material handling) pool rate=IF THEN ELSE(total

batch-level-1 (material handling) costN0, total batch-level-1(material handling) cost/Total Consumption Batch-Level-1 CostPool Activity Driver (Number of Moves), 20)Units: Dollar/Move

12. batch-level-1 activity cost driver (number of moves) consump-tion ratio by Basic model=Basic Model Production Rate/batchsize of Basic modelbatch-level-1 activity cost driver consump-tion ratio by each batch of Basic modelUnits: Move/Month

13. batch-level-1 activity cost driver (number of moves) consump-tion ratio by Deluxe model=Deluxe Model Production Rate/batch size of Deluxe modelbatch-level-1 activity cost driverconsumption ratio by each batch of Deluxe modelUnits:Move/Month

14. batch-level-1 activity cost driver consumption ratio by eachbatch of Basic model=100Units: Move/Batch

15. batch-level-1 activity cost driver consumption ratio by eachbatch of Deluxe model=66.67Units: Move/Batch

16. Batch-Level-1 Cost (Material Handling) Rate for Basic Mod-el=batch-level-1 activity cost driver (number of moves)consumption ratio by Basic modelbatch-level-1(materialhandling) pool rateUnits: Dollar/Month

17. batch-level-2 (equipments setup) pool rate=IF THEN ELSE(total batch-level-2 (equipments setup) costN0, total batch-level-2 (equipments setup) cost /Total Consumption Batch-Level-2 Cost Pool Activity Driver (Number of Setups), 1200)Units: Dollar/Setup

18. batch-level-2 activity cost driver (number of setups) con-sumption ratio by Basic model=Basic Model Production Rate/batch size of Basic model batch-level-2 activity cost driverconsumption ratio by each batch of Basic modelUnits:Setup/Month

19. batch-level-2 activity cost driver (number of setups) consumptionratio by Deluxe model=Deluxe Model Production Rate/batch sizeof Deluxe modelbatch-level-2 activity cost driver consumptionratio by each batch of Deluxe modelUnits: Setup/Month

20. batch-level-2 activity cost driver consumption ratio by eachbatch of Basic model=1Units: Setup/Batch

21. batch-level-2 activity cost driver consumption ratio by eachbatch of Deluxe model=1Units: Setup/Batch

22. Batch-Level-2 Cost (Equipments Setup) Rate for Basic Mod-el=batch-level-2 (equipments setup) pool rate batch-level-2 activity cost driver (number of setups) consumptionratio by Basic modelUnits: Dollar/Month

23. Batch-Level-2 Cost (Equipments Setup) Rate for DeluxeModel= batch-level-2 (equipments setup) pool rate -batch-level-2 activity cost driver (number of setups) consump-tion ratio by Deluxe modelUnits: Dollar/Month

24. Batch-Level-l Cost (Material Handling) Rate for Deluxe Mod-el=batch-level-1 (material handling) pool rate batch-level-1 activity cost driver (number of moves) consumptionratio by Deluxe modelUnits: Dollar/Month

25. Deluxe model cost of goods manufactured=total overhead costfor Deluxe model+Total Prime Cost of Deluxe Model+TotalHolding Cost of Deluxe ModelUnits: Dollar

26. Deluxe model cost of goods manufactured per unit=IF THENELSE(Total Production of Deluxe ModelN0,Deluxe model cost ofgoods manufactured/Total Production of Deluxe Model, 181.21)Units: Dollar/Unit

27. Deluxe model holding cost coefcient=0.05Units: 1/Month

28. Deluxe model markup=0.9866Units: Dmnl29. Deluxe model order fulllment rates:=GET XLS DATA(FD.xls,Sheet1, 1, B3) -Units: Order/Month

30. Deluxe Model Production Rate:=GET XLS DATA(PD.xls,Sheet1, 1, B3)Units: Unit/Month

31. Deluxe model selling price=(1+Deluxe model markup)De-luxe model cost of goods manufactured per unitUnits: Dollar/Unit

32. Deluxe Model Shipping Rate:=GET XLS DATA(SD.xls, Sheet1,1, B3)Units: Unit/Month

33. Facility-Level (Providing Space) Cost Rate for Basic Model=-facility-level (providing space) pool ratefacility-level activitycost driver (machining hours) consumption ratio by BasicmodelUnits: Dollar/Month

34. Facility-Level (Providing Space) Cost Rate for Deluxe Mod-el=facility-level (providing space) pool ratefacility-levelactivity cost driver (machining hours) consumption ratio byDeluxe modelUnits: Dollar/Month

35. facility-level (providing space) pool rate=IF THEN ELSE(totalcost of facility-level (providing space) cost poolN0, total costof facility-level (providing space) cost pool/Total ConsumptionFacility-Level Cost Pool Activity Driver (Machining Hours), 2)Units: Dollar/MachiningHr

36. facility-level activity cost driver (machining hours) consump-tion ratio by Basic model=Basic Model Shipping Rate-facility-level activity cost driver consumption ratio by eachunit of Basic modelUnits: MachiningHr/Month

37. facility-level activity cost driver (machining hours) consump-tion ratio by Deluxe model=Deluxe Model Shipping Rate-facility-level activity cost driver consumption ratio by each unitof Deluxe modelUnits: MachiningHr/Month

38. facility-level activity cost driver consumption ratio by each unitof Basic model=0.25Units: MachiningHr/Unit

39. facility-level activity cost driver consumption ratio by each unitof Deluxe model=0.5 -Units: MachiningHr/Unit

40. FINAL TIME=12Units: Month41. fractional prime cost ratio for Basicmodel=80Units: Dollar/Unit42. fractional prime cost ratio for Deluxe model=160Units:

Dollar/Unit43. Holdeing Cost Rate of Deluxe Model=holding cost fractional

ratio for Deluxe model Inventory Level of Deluxe ModelUnits: Dollar/Month

44. holding cost fractional ratio for Basic model=Basic modelholding cost coefcientBasic model cost of goods manufac-tured per unitUnits: Dollar/(UnitMonth)

45. holding cost fractional ratio for Deluxe model=Deluxe modelholding cost coefcientDeluxe model cost of goods manufac-tured per unitUnits: Dollar/(MonthUnit)

46. Holding Cost Rate of Basic Model=holding cost fractional ratiofor Basic model Inventory Level of Basic ModelUnits: Dollar/Month

47. INITIAL TIME=1 Units: Month48. Inventory Level of Basic Model=INTEG (Basic Model Production

Rate-Basic Model Shipping Rate,0)Units: Unit49. Inventory Level of Deluxe Model=INTEG (Deluxe Model

Production Rate-Deluxe Model Shipping Rate,0)Units:Unit

50. Prime Cost Rate of Basic Model=Basic Model ProductionRate fractional prime cost ratio for Basic modelUnits:Dollar/Month

51. Prime Cost Rate of Deluxe Model=Deluxe Model ProductionRate fractional prime cost ratio for Deluxe modelUnits:Dollar/Month

52. order-level-1 activity cost driver (number of orders) consump-tion ratio by Basic model=Basic model order fulllmentratesorder-level-1 activity cost driver consumption ratio by

each order of Basic modelUnits: Order/Month

155A.H. Khataie et al. / Decision Support Systems 52 (2011) 14215653. order-level-1 activity cost driver (number of orders) consump-tion ratio by Deluxe model=Deluxe model order fulllmentratesorder-level-1 activity cost driver consumption ratio byeach order of Deluxe modelUnits: Order/Month

54. order-level-1 activity cost driver consumption ratio by eachorder of Basic model=1Units: Dmnl

55. order-level-1 activity cost driver consumption ratio by eachorder of Deluxe model=1Units: Dmnl

56. Order-level-1 Cost Rate for Basic Model=order-level-1 poolrateorder-level-1 activity cost driver (number of orders)consumption ratio by Basic modelUnits: Dollar/Month

57. Order-level-1 Cost Rate for Deluxe Model=order-level-1activity cost driver (number of orders) consumption ratio byDeluxe modelorder-level-1 pool rateUnits: Dollar/Month

58. order-level-1 pool rate=IF THEN ELSE(total cost of order-level-1 cost poolN0,total cost of order-level-1 cost pool/TotalConsumption Order-level-1 Cost Pool Activity Driver (Number ofOrders), 173.33)Units: Dollar/Order

59. order-level-2 activity cost driver (maintenances hours) con-sumption ratio by Basic model=Basic model order fulllmentratesorder-level-2 activity cost driver consumption ratio byeach order of Basic modelUnits: MaintenanceHr/Month

60. order-level-2 activity cost driver (maintenances hours) con-sumption ratio by Deluxe model=order-level-2 activity costdriver consumption ratio by each order of Deluxe modelDeluxemodel order fulllment ratesUnits: MaintenanceHr/Month

61. order-level-2 activity cost driver consumption ratio by eachorder of Basic model=4Units: MaintenanceHr/Order

62. order-level-2 activity cost driver consumption ratio by eachorder of Deluxe model=6Units: MaintenanceHr/Order

63. Order-level-2 Cost Rate for BasicModel=order-level-2 activitycost driver (maintenances hours) consumption ratio by Basicmodelorder-level-2 pool rateUnits: Dollar/Month

64. Order-level-2 Cost Rate for Deluxe Model=order-level-2activity cost driver (maintenances hours) consumption ratio byDeluxe modelorder-level-2 pool rateUnits: Dollar/Month

65. order-level-2 pool rate=IF THEN ELSE(total cost of order-level-2 cost poolN0, total cost of order-level-2 cost pool/To-tal Consumption Order-level-2 Cost Pool Activity Driver (Main-tenances Hours), 58.5)Units: Dollar/MaintenanceHr

66. SAVEPER=TIME STEPUnits: Month (The frequency withwhich output is stored.)

67. TIME STEP=1 Units: Month (The time step for the simulation.)68. Total Batch-Level-1 (Material Handling) Cost for Basic Mod-

el=INTEG (Batch-Level-1 Cost (Material Handling) Rate forBasic Model,0)Units: Dollar

69. Total Batch-Level-1 (Material Handling) Cost for DeluxeModel=INTEG (Batch-Level-l Cost (Material Handling) Ratefor Deluxe Model,0)Units: Dollar

70. total batch-level-1 (material handling) cost=Total Batch-Level-1 (Material Handling) Cost for Basic Model+TotalBatch-Level-1 (Material Handling) Cost for Deluxe Model-Units: Dollar

71. Total Batch-Level-2 (Equipments Setup) Cost for Basic Mod-el=INTEG (Batch-Level-2 Cost (Equipments Setup) Rate forBasic Model,0)Units: Dollar

72. Total Batch-Level-2 (Equipments Setup) Cost for DeluxeModel=INTEG (Batch-Level-2 Cost (Equipments Setup) Ratefor Deluxe Model,0)Units: Dollar

73. total batch-level-2 (equipments setup) cost=Total Batch-Level-2 (Equipments Setup) Cost for Basic Model+TotalBatch-Level-2 (Equipments Setup) Cost for Deluxe ModelU-nits: Dollar

74. Total Consumption Batch-Level-1 Cost Pool Activity Driver (Num-ber ofMoves)=INTEG (Total Consumption Ratio of Batch-Level-1

Activity Cost Driver (Number of Moves), 1)Units: Move75. Total Consumption Batch-Level-2 Cost Pool Activity Driver (Num-ber of Setups)=INTEG (Total ConsumptionRatio of Batch-Level-2Activity Cost Driver (Number of Setups), 1)Units: Setup

76. Total Consumption Facility-Level Cost Pool Activity Driver(Machining Hours)=INTEG (Total Consumption Ratio ofFacility-Level Activity Cost Driver (Machining Hours), 1)Units: MachiningHr

77. Total Consumption Order-level-1 Cost Pool Activity Driver(Number of Orders)=INTEG (Total Consumption Ratio ofOrder-level-1 Activity Cost Driver (Number of Orderss), 1)Units: Order

78. Total Consumption Order-level-2 Cost Pool Activity Driver(Maintenances Hours)=INTEG (Total Consumption Ratio ofOrder-level-2 Activity Cost Driver (Maintenances Hours), 1)Units: MaintenanceHr

79. Total Consumption Ratio of Batch-Level-1 Activity Cost Driver(Number of Moves)=batch-level-1 activity cost driver (num-ber of moves) consumption ratio by Basic model+batch-level-1 activity cost driver (number of moves) consumptionratio by Deluxe modelUnits: Move/Month

80. Total Consumption Ratio of Batch-Level-2 Activity Cost Driver(Number of Setups)=batch-level-2 activity cost driver (num-ber of setups) consumption ratio by Basic model+batch-level-2 activity cost driver (number of setups) consumptionratio by Deluxe modelUnits: Setup/Month

81. Total Consumption Ratio of Facility-Level Activity Cost Driver(Machining Hours)=facility-level activity cost driver (ma-chining hours) consumption ratio by Basic model+facility-level activity cost driver (machining hours) consumption ratioby Deluxe modelUnits: MachiningHr/Month

82. Total Consumption Ratio of Order-level-1 Activity Cost Driver(Number of Orderss)= order-level-1 activity cost driver(number of orders) consumption ratio by Basic model+or-der-level-1 activity cost driver (number of orders) consumptionratio by Deluxe modelUnits: Order/Month

83. Total Consumption Ratio of Order-level-2 Activity Cost Driver(Maintenances Hours)=order-level-2 activity cost driver(maintenances hours) consumption ratio by Basic model+or-der-level-2 activity cost driver (maintenances hours) consump-tion ratio by Deluxe modelUnits: MaintenanceHr/Month

84. total cost of facility-level (providing space) cost pool=TotalFacility-Level (Providing Space) Cost for Basic Model+TotalFacility-Level (Providing Space) Cost for Deluxe ModelUnits:Dollar

85. total cost of order-level-1 cost pool=Total Order-level-1 Costfor Basic Model+Total Order-level-1 Cost for Deluxe Mod-elUnits: Dollar

86. total cost of order-level-2 cost pool=Total Order-level-2 Costfor Basic Model+Total Order-level-2 Cost for Deluxe Mod-elUnits: Dollar

87. Total Facility-Level (Providing Space) Cost for Basic Mode-l=INTEG (Facility-Level (Providing Space) Cost Rate for BasicModel,0)Units: Dollar

88. Total Facility-Level (Providing Space) Cost for Deluxe Mod-el=INTEG (Facility-Level (Providing Space) Cost Rate forDeluxe Model,0)Units: Dollar

89. Total Holding Cost of Basic Model=INTEG (Holding Cost Rate ofBasic Model,0)Units: Dollar

90. Total Holding Cost of DeluxeModel=INTEG (Holdeing Cost Rateof Deluxe Model,0)Units: Dollar

91. total overhead cost for Basic model=Total Batch-Level-1(Material Handling) Cost for Basic Model+ Total Batch-Level-2 (Equipments Setup) Cost for Basic Model+TotalFacility-Level (Providing Space) Cost for Basic Model+TotalOrder-level-1 Cost for Basic Model+Total Order-level-2 Cost

for Basic ModelUnits: Dollar

92. total overhead cost for Deluxe model=Total Batch-Level-1(Material Handling) Cost for Deluxe Model+Total Batch-Level-2 (Equipments Setup) Cost for Deluxe Model+TotalFacility-Level (Providing Space) Cost for Deluxe Model+TotalOrder-level-1 Cost for Deluxe Model+Total Order-level-2Cost for Deluxe ModelUnits: Dollar

93. Total Prime Cost of Basic Model=INTEG (Prime Cost Rate ofBasic Model,0)Units: Dollar

94. Total Prime Cost of Deluxe Model=INTEG (Prime Cost Rate ofDeluxe Model,0)Units: Dollar

95. Total Order-level-1 Cost for Basic Model=INTEG (Order-

Studies Journal 10 (2006) 4977.[5] D. Cagwin, M.J. Bouwman, The association between activity-based costing and

[13] R.S. Kaplan, S.R. Anderson, Time-driven activity-based costing, Harvard BusinessReview 82 (2004) 131138.

[14] R. Kee, Integrating activity-based costing with theory of constraints to enhanceproduction-related decision-making, Accounting Horizons 9 (1995) 4861.

[15] T. Kennedy, J. Afeck-Graves, The impact of activity-based costing techniques onrm performance, Journal of Management Accounting Research 13 (2001) 1945.

[16] A. Khataie, F.M. Defersha, A.A. Bulgak, Amulti-objective optimization approach fororder management: incorporating Activity-Based Costing in supply chains,International Journal of Production Research 48 (2010) 50075020.

[17] E.T. Kirche, S.N. Kadipasaoglu, B.M. Khumawala, Maximizing supply chain protswith effective order management: integration of activity-based costing andtheory of constraints with mixed-integer modeling, International Journal ofProduction Research 43 (2005) 12971311.

[18] E. Kirche, R. Srivastava, An ABC-based cost model with inventory and order level

156 A.H. Khataie et al. / Decision Support Systems 52 (2011) 142156improvement in nancial performance, Management Accounting Research 13(2002) 139.

[6] R. Cooper, R.S. Kaplan, The Design of Cost Management Systems: Text, Cases andReadings, First ed. Prentice-Hall, Englewood Cliffs, 1991.

[7] M. Gosselin, The effect of strategy and organizational structure on the adoptionand implementation of activity-based costing, Accounting, Organizations andSociety 22 (1997) 105122.

[8] M. Gosselin, A review of activity-based costing: technique, implementation, andconsequences, Handbook of Management Accounting Research 2 (2007)641671.

[9] A. Gregoriades, B. Karakostas, Unifying business objects and system dynamics as aparadigm for developing decision support systems, Decision Support Systems 37(2004) 307311.

[10] D.R. Hansen, M.M. Mowen, N.S. Elias, D.W. Senkow, Management Accounting,Canadian, Fifth ed. Nelson, Canada, 2001.

[11] J.S. Holmen, ABC vs. TOC: it's a matter of time, Management Accounting 76 (1995)3740.

[12] C.D. Ittner, W.N. Lanen, D.F. Larcker, The association between activity-basedcosting and manufacturing performance, Journal of Accounting Research 40(2002) 711726.level-1 Cost Rate for Basic Model,0)Units: Dollar96. Total Order-level-1 Cost for Deluxe Model=INTEG (Order-

level-1 Cost Rate for Deluxe Model,0)Units: Dollar97. Total Order-level-2 Cost for Basic Model=INTEG (Order-

level-2 Cost Rate for Basic Model,0)Units: Dollar98. Total Order-level-2 Cost for Deluxe Model=INTEG (Order-

level-2 Cost Rate for Deluxe Model,0)Units: Dollar99. Total Production of Basic Model=INTEG (Basic Model Produc-

tion Rate,0)Units: Unit100. Total Production of Deluxe Model=INTEG (Deluxe Model

Production Rate, 0) - Units: Unit101. Total Sales of Basic Model=INTEG (Basic Model Shipping

Rate,1)Units: Unit102. Total Sales of Deluxe Model=INTEG (Deluxe Model Shipping

Rate,1)Units: Unit

References

[1] T.K. Abdel-Hamid, S.E. Madnick, On the portability of quantitative softwareestimation models, Inf and Management 13 (1987) 110.

[2] B.J. Angerhofer, M.C. Angelides, System Dynamics Modeling in Supply ChainManagement: Research Review, Proceedings of the 2000 Winter SimulationConference, 2000, pp. 342351.

[3] T.S. Baines, D.K. Harrison, An opportunity for system dynamics in manufacturingsystem modeling, Production Planning and Control 10 (1999) 542552.

[4] D. Cagwin, K.J. Barker, Activity-based costing, total quality management andbusiness process reengineering: their separate and concurrent association withimprovement in nancial performance, Academy of Accounting and Financialcosts: a comparison with TOC, International Journal of Production Research 43(2005) 16851710.

[19] D. Li, C. O'Brien, Integrated decision modeling of supply chain efciency,International Journal of Production Economics 59 (1999) 147157.

[20] D. Li, C. O'Brien, A quantitative analysis of relationships between product typesand supply chain strategies, International Journal of Production Economics 73(2001) 2939.

[21] J. Macedo, R. Ruiz Usano, J. Framinan Torres, A real time cost monitoring system forbusiness process reengineering, Proceedings from IFACWorkshop (1997) 411416.

[22] S.A. Malik, W.G. Sullivan, Impact of ABC information on product mix and costingdecisions, IEEE Transactions on Engineering Management 42 (1995) 171176.

[23] A.C. Marquez, C. Blanchar, A Decision Support System for evaluating operationsinvestments in high-technology business, Decision Support Systems 41 (2006)472487.

[24] C. Martinez-Olvera, Benets of using hybrid business models within a supplychain, International Journal of Production Economics 120 (2009) 501511.

[25] B. Novievi, L. Anti, Total quality management and activity-based costing,Economics and Organization 1 (1999) 18.

[26] W. Sen, S. Pokharel, W.Y. Lei, Supply chain positioning strategy integration,evaluation, simulation, and optimization, Computers and Industrial Engineering46 (2004) 781792.

[27] C. Sheu, M.H. Chen, S. Kovar, Integrating ABC and TOC for better manufacturingdecision making, Integrated Manufacturing System 14 (2003) 433441.

[28] J.D. Sterman, Business Dynamics: Systems Thinking and Modeling for a ComplexWorld, First ed. Irwin, CF: McGraw-Hill, 2000.

Amir H. Khataie obtained his B.Sc. degree in Industrial Engineering in 2005 from AmirKabir University of Technology (Tehran Polytechnic) consequently; he received M. Eng.degree in Engineering Management from the University of Ottawa in August 2007.Currently he is a part-time faculty and PhD candidate in the department of Mechanicaland Industrial Engineering at Concordia University. Amir H. Khataie researchpreferences are within hybrid hard and soft OR modeling approach, supply chainmanagement, activity-based costing. His current research is application of accountingapproaches in developing supply chain cost management Decision Support Systems.

Dr. Akif A. Bulgak is a professor at the Department of Mechanical and IndustrialEngineering at Concordia University. He obtained his B.Sc. degree from IstanbulTechnical University in Mechanical/Industrial Engineering and his M.Sc. and Ph.D.degrees from the University of Wisconsin-Madison, USA, all in Industrial Engineering.Dr. Bulgak's research areas include Modeling, Performance Evaluation, DesignOptimization, and Economics of Flexible Manufacturing/Assembly Systems, StochasticOptimization, and Revenue Management, and Supply Chain Management. Dr. Bulgak isa registered professional engineer at Professional Engineers Ontario.

Dr. Juan J. Segovia is currently an Associate Professor of Accountancy in the JohnMolson School of Business, Concordia University, Montreal, Canada. He obtained hisDoctorate in Business Administration in 1979 from the University of Paris-Dauphine,France. From the same university, he obtained the Diplme d'tudes Approfondies:Business Administration (DEA). Professor Segovia obtained the degree of Bachelor ofCommerceMajor: Accounting, from the Universidad de Guanajuato, Mexico. His areaof research includes Accounting Education and Management Accounting. His papershave been published in various prestigious journals, e.g., The British AccountingReview, The Accounting Educators' Journal, CMA Magazine, Journal of Business andBehavioural Sciences. In addition, he has presented at several conferences both nationaland international. His research interests are in the areas of Accounting Education,Activity-Based Costing (ABC), Activity-Based Management (ABM), PerformanceMeasurements, and Business Strategy.

Activity-Based Costing and Management applied in a hybrid Decision Support System for order management1. Introduction2. Background3. General illustrative problem4. Order Management MIP decision support model5. Numerical example6. System Dynamics decision support model6.1. First part6.2. Second to sixth part6.3. Seventh part

7. Conclusion and future workAppendixReferences

activity based costing and management applied in a hybrid decision support system for order...

Documents

moreain order management

management approach

ptp management needs

ptp approach

cost objects

cost control1

management decisionprotable

specic cost object