demand forecating,final

Upload: erjaspreet

Post on 30-May-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/14/2019 Demand Forecating,Final

    1/32

    DEMAND FORECASTING

    Talwinder Singh

    Roll No.80782011

  • 8/14/2019 Demand Forecating,Final

    2/32

    .

    Before making investment in men, machinery, toolingetc.,organization has to answer certain questions. Fore.g. it must knows in advance:

    3) What is the capacity of company like Power plant200MW?

    4) What is the amount of capital to carry out forproduction?

    5) What is the type of production system the companyshould follow?

    6) What is the level of inventories that have to kept atdifferent intervals of time?

  • 8/14/2019 Demand Forecating,Final

    3/32

    Definitions

    Demand forecasting means to look into future, whatproducts do the customer prefer, in what quantity andby what time or during which period.

    Forecasting is the process of estimating future eventsby casting forward past data, the past data issystematically combined in a predetermined way toobtain the estimate of the future.

  • 8/14/2019 Demand Forecating,Final

    4/32

    Scope of forecasting

    Planning activities

    1) Capacity planning, layout decisions.

    2) Process design3) Strategy for marketing

    4) Budget

  • 8/14/2019 Demand Forecating,Final

    5/32

    .

    Scheduling / Controlling activities

    1) Aggregate planning2) Master production schedule

    3) Production programme,Prodution rate, Lot size etc.

    4) Man power requirement planning

    5) Material requirement planning6) Inventory management and control

  • 8/14/2019 Demand Forecating,Final

    6/32

    Difference between forecasting andprediction

    Method

    Personnel bias Reproducibility

    Error analysis

  • 8/14/2019 Demand Forecating,Final

    7/32

    Cost of forecasting

    Costs of increased activity

    Opportunity costs

    Opportunity cost

    Cost of increased activity

    Total cost

    Forecasting activity

    Accuracy

    Cos

    tofforecasting

    Judgment method

    Time series

    method

    Causal method

  • 8/14/2019 Demand Forecating,Final

    8/32

    Types of forecasting techniques

    Qualitative techniques(Judgment methods)

    - Salesforce estimate- Experts opinion

    - Market research

    - Panel consensus- Delphi method

  • 8/14/2019 Demand Forecating,Final

    9/32

    Quantitative techniques

    (a) Causal methods

    - Correlation and regression analysis

    (b) Time series analysis

    - Simple moving averages

    - Weighted moving averages

    - Exponential smoothing forecasting

  • 8/14/2019 Demand Forecating,Final

    10/32

    Type of forecasting on the basis oftime horizon

    Short range forecasting- Up to 3 months into the future

    Medium range forecasting- 3 months 2 years into the future

    Long range forecasting

    - > 2 years

  • 8/14/2019 Demand Forecating,Final

    11/32

    Qualitative techniques(Judgment methods)

    Past sales data is not available

    Forecasting has to be done for a verydistinct period (say for 5 year)

  • 8/14/2019 Demand Forecating,Final

    12/32

    Salesforce estimates

    Also called Grass root method

    Employees from low level like salesrepresentative, marketing people.

    Forecast errors occur due topersonnel bias, incentive schemes.

  • 8/14/2019 Demand Forecating,Final

    13/32

    Executive opinion method

    Managerial level of people

    Costly Forecast errors occurs if managers

    changes the forecast figure

  • 8/14/2019 Demand Forecating,Final

    14/32

    Panel consensus

    Very popular method

    People from all level of management

    Open meeting

    Everybody comes turn by turn, with opinionand why they feel in such a way

    At the end of meeting common figure

    comes out with the consensus of allmembers

    Forecast error occurs due to organizationculture

  • 8/14/2019 Demand Forecating,Final

    15/32

    Delphi method

    Introduce to overcome the limitations of panelconsensus method

    Group of experts Their identity is concealed

    All work in written form and not a open meeting Coordinator / moderator Questionnaire Summary report

    Again revised by experts Generally in 3 rounds, we will get the forecast figure

  • 8/14/2019 Demand Forecating,Final

    16/32

    Quantitative techniques

    (a) Causal methods- These methods are based on cause and effect relationship

    - Relationship between 2 or more factors

    - Dependent and independent variables

    - Dependent variables like demand, sales etc.- Independent variables like population, advertisement expenditures etc.

    Regression line , y=a + bx

    Depende

    ntvariable

    Independent variable

  • 8/14/2019 Demand Forecating,Final

    17/32

    Example

  • 8/14/2019 Demand Forecating,Final

    18/32

    .

  • 8/14/2019 Demand Forecating,Final

    19/32

  • 8/14/2019 Demand Forecating,Final

    20/32

    Time series methods

    Driving a car by seeing rear view

    Simple moving averages

    Weighted moving averages Exponential smoothing

  • 8/14/2019 Demand Forecating,Final

    21/32

    Simple moving average

  • 8/14/2019 Demand Forecating,Final

    22/32

    Weighted moving average

  • 8/14/2019 Demand Forecating,Final

    23/32

    Example

  • 8/14/2019 Demand Forecating,Final

    24/32

    Solution

  • 8/14/2019 Demand Forecating,Final

    25/32

    Exponential smoothing method

    One of the disadvantages of the movingaverage forecasting is the laboriousoperation of maintaining the data for allprevious years.

    This method gives more weightage torecent demand and the weight assigned toolder periods decreases exponentially.

    This method requires:

    - most recent actual demand- most recent forecast- smoothing constant ()

  • 8/14/2019 Demand Forecating,Final

    26/32

    Exponential smoothing method

    Ft+1 = Dt + (1- )Ft It is a modified form of weighted average method Ft+1 = Dt + (1- ){ Dt-1+ (1- )Ft-1}

    = Dt + (1- ) Dt-1+ (1- )2 Dt-2 + ..}

    The weights are : (1- )0

    (1- )1

    (1- )2 and so on. And this decay in weightages is in exponential

    manner.

  • 8/14/2019 Demand Forecating,Final

    27/32

  • 8/14/2019 Demand Forecating,Final

    28/32

    Impact of n and

    n low

    - more fluctuating

    - but there is a closer following of the trend

    If we want to give more weightage to recent demand ascompared to old data, take large value of

    If demand is stable , take small value of

  • 8/14/2019 Demand Forecating,Final

    29/32

    Multiplicative seasonal time seriesmethod

    Find out average demand per year

    Find out the parameter / quantity

    called seasonal index (S.I.) S.I. = Actual demand during a season / Average demandduring a season

    Calculate average seasonal index for

    a season (A.S.I.)(A.S.I.)Q1 = {(S.I.Q1)y1 + (S.I.Q1)y2 + (S.I.Q1)y3 +

    (S.I.Q1)y4} / 4

  • 8/14/2019 Demand Forecating,Final

    30/32

    Example Find out quarterly forecast for the number of products expected in

    the next year. The quarterly demand for the past four years is shown

    in table. The total forecasted demand for the next year be 2600units.

    215285170100Q41160830590520Q3

    725585370335Q2

    1001007045Q1

    Y4Y3Y2Y1

    S l ti

  • 8/14/2019 Demand Forecating,Final

    31/32

    Solution

    Step 1. Calculate average demand

    per year.for example, year 1 = 45 + 335 + 520 +100) /4 = 250

    year 2 = 300, year 3 = 450, year 4 =550

    Step 2. Find S.I., like 45/250 =0.18 Step 3. Find A.S.I.

    A.S.I.(Q1) = (0.18 + 0.23 + 0.22 +0.18 ) /4 = 0.2

    A.S.I.(Q2) = 1.3A.S.I.(Q3) = 2.0A.S.I.(Q4) = 0.5So, forecast for the next yearQ1 = 0.2 x (2600/4) = 0.2 x 650 = 130

    units

    Q2 = 1.3 x (2600/4) = 1.3 x 650 = 845units

    Q3 = 2.0 x (2600/4) = 2.0 x 650 = 1300units

    Q4 = 0.5 x (2600/4) = 0.5 x 650 = 325units

    0.390.630.560.4Q4

    2.111.841.962.08Q3

    1.321.31.231.34Q2

    0.180.220.230.18Q1

    Y4Y3Y2Y1

  • 8/14/2019 Demand Forecating,Final

    32/32

    Thanks!