A Variable Neighbourhood Search Algorithm for the Beam Angle Selection Problem in Radiation Therapy

Intensity modulated radiation therapy (IMRT) is one of the most common techniques in radiation therapy for cancer treatment. One first problem within IMRT is to find the best possible combination of beam angles so we can deliver a clinically acceptable treatment plan to the patient. This problem, also known as the Beam Angle Optimization (BAO) problem, is very hard to solve and, thus, different strategies combining heuristics and mathematical programming have been proposed in the literature. In this paper, we propose to combine a simple yet efficient variable neighborhood search (VNS) algorithm with an interior point strategy to (approximately) solve the BAO problem. We apply our hybrid algorithm on a prostate case and compare the obtained results to those obtained by a previously proposed local search algorithm. Results show that the VNS is able to find beam angles configurations that are quite competitive w.r.t. those found by the local search algorithm.