Designing Less Myopic Routing Policies with Genetic Programming for the Electric Vehicle Routing Problem

Created by W.Langdon from gp-bibliography.bib Revision:1.8834

@InProceedings{Durasevic:2024:EA2,

• author = "Marko Durasevic and Francisco Javier {Gil Gala}",
• title = "Designing Less Myopic Routing Policies with Genetic Programming for the Electric Vehicle Routing Problem",
• booktitle = "Artificial Evolution, EA-2024",
• year = "2024",
• editor = "Pierrick Legrand",
• volume = "15926",
• series = "LNCS",
• address = "Bordeaux, France",
• month = oct # " 29-31",
• keywords = "genetic algorithms, genetic programming, Electric vehicle routing problem, Routing rule: Poster",
• isbn13 = "978-3-032-07998-5",
• URL = "http://evr.zemris.fer.hr/EN_Home.html",
• URL = "https://digibuo.uniovi.es/dspace/handle/10651/83074",
• URL = "https://digibuo.uniovi.es/dspace/bitstream/handle/10651/83074/Designing%20Less%20Myopic%20Routing%20Policies%20with%20Genetic%20Programming%20for%20the%20Electric%20Vehicle%20Routing%20Problem.pdf",
• size = "8 pages",
• abstract = "Automatically designing heuristics with genetic programming (GP) has become an increasingly researched topic in the last several years. Recently, this methodology has also been applied for designing routing policies (RPs) for the electric vehicle routing problem (EVRP). RPs are simple heuristics that construct the solution in a way that each time a new customer needs to be visited, the RP selects the customer based on certain problem characteristics. However, the performance of such a methodology is significantly influenced by the problem characteristics that can be used to construct the heuristics. Therefore, if GP does no have access to all the relevant characteristics, the heuristics which it will design will be myopic and the solutions they construct will be of poor quality. Because of this it is important to select the appropriate properties that GP will use to construct new heuristics. Therefore, the goal of this study is to analyse how including additional characteristics that are connected to the position of vehicles can improve the quality of RPs generated by GP. The results demonstrate that for certain problems this newly included information can significantly improve the performance of the generated heuristics.",
• notes = "EA2024",

}

Genetic Programming entries for Marko Durasevic Francisco Javier Gil Gala

Citations