README.md

EuclideanIRP

Solves Inventory Routing Problem (IRP) on Euclidean Instances

Dependencies

dapcstp

Concorde

Gurobi

Usage

Build the source files using make. After building the executable of the dapcstp solver, place the executable file in each of the following folders: EuclideanIRPdeleteAddLS, EuclideanIRPaddLS, EuclideanIRPdeleteLS, EuclideanIRPgreedy, EuclideanIRPpd. Each heuristic is currently set to run on one instance from EuclideanIRPinstances/inputsIterNmult100Htimes100_260. To obtain the output, create an outputs directory outputsIterNmult100Htimes100_260 in the directory of the heuristic called. Each of the source files may be modified to run over all instances in any instance directory by adding a loop over the parameter (H,N, or T) that is varying and a loop over the instance number from 1 to 100.

1. In the EuclideanIRPlp directory, solveMetricIPscf8coresCutOffLBToursSoln3displayrhCC.cpp models a given IRP instance as a MIP and calls Gurobi to find a lower bound on the optimal cost within 10% MIPgap.
2. In the EuclideanIRPdeleteAddLS directory, EuclideanBestPrioritizedOperationsLSwCutOffInitializeAddSolnToursSoln3cc.cpp finds a feasible solution using local search applying DELETEs, ADDs, and pairwise DELETE-ADDs.
3. In the EuclideanIRPaddLS directory, EuclideanBestAddLSwCutOffToursSolnForServer6cc.cpp finds a feasible solution using local search applying ADDs.
4. In the EuclideanIRPdeleteLS directory, EuclideanBestDeleteLS4wCutOffToursSolnCC.cpp finds a feasible solution using local search applying DELETEs.
5. In the EuclideanIRPgreedy directory, EuclideanGreedy7PrioritizedOperationsCutOffToursSolnCC.cpp finds a feasible solution by greedily chosing low density set covers to serve demands.
6. In the EuclideanIRPpd directory, EuclideanPrimalDualPrimalOnlyPrioritizedOperationsCutOffToursSolnCC.cpp finds a feasible solution via a primal dual framework to determine visits.