Kernel Search for Overhead Wires Network Planning for Battery-assisted Trolleybuses

Abstract Battery-assisted trolleybus technology can be seen as a perspective solution to develop public transport in more environmental-friendly way. This paper deals with the heuristic approach to solve the problem of creating minimal required overhead wires network for battery-assisted trolleybuses. A mathematical programming model of the problem is presented. The problem is solved using the Kernel Search method. Kernel Search was historically proven to produce high-quality solutions in many optimization problems. We present results for multiple instances, starting with small problems up to real full-scale Žilina city network. Testing data was generated using simulation software (OptSim) from vehicle schedules provided by the public transport company in Žilina (DPMŽ) and map data from OpenStreetMap project. Heuristic results are compared with near-optimum results obtained by a mathematical programming solver.