Simulation Analysis of Simulation Analysis of Truck Driver Scheduling Truck Driver Scheduling
RulesRules
Eric C. ErvinRussell C. Harris
J.B. Hunt Transport, Inc.615 J.B. Hunt Corporate Drive
P.O. Box 130Lowell, Arkansas 72745, U.S.A.
Presented By: Craig Rachel, Midwestern State University
Overview:Overview:
IntroductionBackgroundObjectiveApproachResults & DiscussionConclusion
IntroductionIntroduction
Redefined Employee Workday Modified Planning Process Productivity Changes that will influence the profitability of
customer contracts
Impacts of Implementation of Scheduling Changes
Change in regulations on hours of service (HOS) for truck drivers
• 120,000 unique freight lanes• 5,000 tractors• 10,000 trailers
IntroductionIntroduction
Order-to-Delivery ProcessTruck Driver’s daily routine
Modeling of Two Important Aspects
• People• Equipment• Material• Information
To show interaction and flow of
BackgroundBackground
First change in 60 years
ObjectiveObjective
Determine the impact of the new 2004 HOS rules as they apply to driver utilization, customer on-time service and the nature of the company’s freight network
Develop a strategy to mitigate any negative impact on utilization and efficiency
ApproachApproach
Demand GenerationCapacity ManagementLoad and Tractor AssignmentDriver Log ManagementTransportation ExecutionCustomer Freight Pick-up and Delivery
Six (6) Major Processes Modeled in the Simulation
Demand GenerationDemand Generation
Generated over 12 month periodRepresented data extracted from 1 full year of
actual historyReflected seasonalityDemand that was not accommodated due to lack
of capacity waited up to 24 hours
Capacity ManagementCapacity Management
Capacity = DriverOn dispatch (in service), not on dispatch
(available), at home (until completion of off-duty time)
Derived from data collected at company warehouses
Load and Tractor AssignmentLoad and Tractor Assignment
Maximize efficiencyMinimize empty milesAvoid customer service failuresAvoid assigning loads to drivers who were due
home soonPreference giving to drivers based on how close
to the load they wereDriver needed sufficient hours
Transportation ExecutionTransportation Execution
Company load history characteristics were usedAssumes average velocity rises as trip continuesConsider congestion urban areas and assume
longer trips utilize expressways
Driver LogsDriver Logs
Off-duty, general off-dutyOff-duty, driver in sleeper berth or at
homeOn-duty, drivingOn-duty, not driving. Loading/unloading
Runtime EnvironmentRuntime Environment
2 GHZ CPU, 1 GB RAMSingle Replication took 4 hours on
averageExperiments took 4 replications or 16
hours to completeSimulated the system for 1 year
Results and DiscussionResults and Discussion
10 hours vs. 11 hours driving per shift15 hours vs. 14 hours on duty per shift8 hours vs. 10 hours break time between
shiftsThe non-consecutive vs. consecutive
nature of on duty time
Analysis of results focused on key differences implemented in 2004:
Results and DiscussionResults and Discussion
Only 3 hours in 14 hour work window to cover inspections, fueling, and loading & unloading
Any delays, erodes the 11 hours available to drive
Biggest Finding: Impact of Consecutive Nature of the way on-duty time is logged
Results and DiscussionResults and Discussion
Speed limit of 62 MPHSpeed limit of 65 MPHSpeed limit of 68 MPHSpeed limit of 70 MPH
Baseline 2003 (old scenario) vs. 2004 HOS changes
Marginal improvement at 70 MPH over 62 MPH
Conclusion:
ConclusionConclusion
Expect miles to drop 2-3%Amount of time to deliver load will go upMPH increase only marginally improves
productivity
Communicate to Drivers:
ConclusionConclusion
Service level impact – decrease 2.4%Evaluate special fees for customers
Communicate to Customers:
ConclusionConclusion
Develop a pricing strategy that factors in loss of miles, etcPrepare for potential capacity decreaseOptimize the 3 hours that drivers have for inspections, etcConsider the 14 hour work window when determining how
to dispatch drivers
The Business:
Evaluating Real World ResultsEvaluating Real World Results
Industry given grace period in 2004Utilization is up in 2004, not comparing the
same numbersCompany policy changed, how did this effect
the numbers?
1st 5 Months of 2004
ReferencesReferences
Jain, S., R.W. Workman, L.M. Collins, E.C. Ervin and A.P. Lathrop, 2001. Development of a High Level Supply Chain Simulation Model
Law, Averill M. and W. David Kelton, 1991. Simulation Modeling & Analysis, 2nd Edition, McGraw-Hill, Inc. USA.
Taylor, G. Don, T.S. Meinert, R.C. Killian and G.L. Whicker, 1999. Development and Analysis of Alternative Dispatching Methods in Truckload Trucking.