Power Forecasting andFleet Location Optimization

American Commercial Barge Line LLC

Gail W. DePuy, G. Don Taylor and Amy Bush

Center for Engineering Logistics & Distribution

CELDi

UL02-02

Two separate problemsProblem 1 – Power Forecasting

Problem DefinitionProblem SettingSolution Approach

Problem 2 – Fleet Location OptimizationProblem DefinitionProblem SettingSolution Approach

Project Status Conclusions & Plans

Poster Overview

Center for Engineering Logistics & Distribution

CELDi

Determine which boat carries each bargeMinimize barge handling costsMinimize barge dwell (idle) costsMinimize sub-contract charges

Understand the trade-off between tows with low handling costs (common destinations) versus tows with low dwell costs (different destinations)

Use a technique that can be used on a real-time basis

Problem 1 – Power ForecastingProblem Description

Center for Engineering Logistics & Distribution

CELDi

Ohio River from Cairo, IL (confluence of Ohio & Mississippi Rivers) to Pittsburgh, PA

Dozens of input/output port locationsHundreds of bargesTwo-week rolling planning horizonDynamic river conditions & stochastic demandVery little help in the published literature

Problem 1 – Power ForecastingProblem Setting

Center for Engineering Logistics & Distribution

CELDi

Simulation Heuristic Extremely detailed analysis Comprehensive & easy to understand & use Useful for comparison & validation Real time barge scheduling capabilities Series of local optima sought; boats can compare the cost of barge

assignment to the cost of assignment to other boats in the system Complex boat management rules including when it is better to turn

the boat Extremely detailed analysis Basic tradeoff between handling & dwell costs Use SIMNET II

Problem 1 – Power ForecastingSolution Approach

Center for Engineering Logistics & Distribution

CELDi

Strategic Issue - Determine river locations to use for various activities (Dynamic Facility Layout Problem) Tow building/Tow breaking Barge Cleaning Barge Repair

Determine number and type of boats needed Locate activities to minimize costs Considerations:

Distances between fleet locations Boat types – operating rules and costs Barge volumes Cleaning and Repair times

Problem 2 – Fleet Location Opt.Problem Description

Center for Engineering Logistics & Distribution

CELDi

New Orleans/Lower Mississippi River 130 current fleet/customer locationsThousands of barges4 boat typesOne month planning horizonDynamic river conditions

Problem 2 – Fleet Location Opt.Problem Setting

Center for Engineering Logistics & Distribution

CELDi

Problem 2 – Fleet Location Opt.Solution Approach

Integer Programming Formulation Objective:

Minimize travel, cleaning, and repair costs Constraints:

Barge volume balance constraints at each location Fleet capacity Boat rules – certain boat types can only operate in

certain areas of river Determination of cleaning locations Determination of repair locations

Center for Engineering Logistics & Distribution

CELDi

Problem 1 – Power Forecasting Basic simulation model has been developed Implementation completed Model enhancements completed Final report submitted

Problem 2 – Fleet Location Optimization IP model developed, verified, and validated Data collection (by sponsor) completed Final report submitted

Project Status

Center for Engineering Logistics & Distribution

CELDi

Two separate deliverables Power Forecasting SIMNET Simulation Tool Fleet Location Optimization Tool (IP program)

Power forecasting SIMNET simulation tool can be used to schedule barges on an hourly basis

Fleet location IP model can be used to perform ‘what-if’ analysis to consider the acquisition of new fleet locations

Concluding Remarks

Center for Engineering Logistics & Distribution

CELDi