The application of genetic algorithm for beam directions optimization in radiotherapy

The genetic algorithm (GA) has been proved to be a useful global optimization tool and has been used to solve many engineering problems. In this paper, a GA with problem-dependant coding scheme is proposed in order to solve the problem of beam direction optimization for radiotherapy. The coding scheme and genetic operations including selection, crossover, mutation and immunity are discussed in detail. In order to speed-up the optimization, some prior expert knowledge is used to guide the genetic process. Finally a simulated case is explored to test the validity of the proposed algorithm for the optimization of beam direction in intensity-modulated radiotherapy planning. The results show that the algorithm is valid and efficient and can find the optimal beam directions within little computation time.