Skip to Main content Skip to Navigation

A Constraint Programming Approach for the Time Dependent Traveling Salesman Problem

Penelope Aguiar-Melgarejo 1
1 M2DisCo - Geometry Processing and Constrained Optimization
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : In the context of urban deliveries, the optimization of delivery tours is usually modeled as a Traveling Salesman Problem (TSP). Side constraints like time-windows constraining the delivery times at the client or breaks for the drivers are also common in this kind of problem and time is an important dimension to take into account to respect these constraints. With travel times’ variability in big cities time also tends to have a greater influence in costs and therefore it should be included in the optimization of delivery routes. The Time-Dependent Traveling Salesman Problem (TDTSP) is the extended version of the Traveling Salesman Problem (TSP) where arc costs depend on the time when the arc is traveled. In this thesis we propose a set of benchmarks for the TDTSP based on real traffic data (obtained from the city of Lyon) and show the interest of handling time dependency in the problem. A study of how to better model time-dependent travel functions in general and specifically for our approach is performed. We introduce and compare different models to solve the TDTSP with Constraint Programming (CP). A first model is derived in a straightforward way from the classical CP model for the TSP. We show that this model is not able to reason on indirect precedence relations, so that it has poor performance on our benchmark. We introduce a new global constraint which is able to exploit indirect precedence relations on time-dependent data, and we introduce a second model which is based on our new constraint. We experimentally compare the two models on our benchmark.
Complete list of metadatas

Cited literature [90 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01514369
Contributor : Équipe Gestionnaire Des Publications Si Liris <>
Submitted on : Wednesday, June 20, 2018 - 11:09:45 AM
Last modification on : Tuesday, November 19, 2019 - 2:41:32 AM
Document(s) archivé(s) le : Monday, September 24, 2018 - 11:54:09 AM

File

thesePenelope.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01514369, version 1

Citation

Penelope Aguiar-Melgarejo. A Constraint Programming Approach for the Time Dependent Traveling Salesman Problem. Computer Science [cs]. INSA Lyon, 2016. English. ⟨NNT : 2016LYSEI142⟩. ⟨hal-01514369⟩

Share

Metrics

Record views

365

Files downloads

1315