Author(s)

K. Kataoka & K. Komaya

Abstract

Crew operation scheduling problem is a large-scale combinatorial problem. Only human experts could make schedule with trial and error in a long time by hand. We propose a new framework for crew operation scheduling using simulated evolution technique, which is used to simulate trial and error approach effectively. Several kinds of knowledge base are defined to raise the effectiveness of the search algorithm to find constraint satisfaction solution. We can find a feasible solution in several minutes with our prototype system, which is developed for the high-frequency subway line. 1 Introduction Since a crew operation scheduling problem is a large-scale combination problem, only human

Keywords