Application of multiobjective genetic algorithms in anatomy based dose optimization in brachytherapy and its comparation with deterministic algorithms

In high dose rate (HDR) brachytherapy the conventional dose optimization algorithms consider the multiple objectives in the form of an aggregate function which combines individual objectives into a single utility value. As a result, the optimization problem becomes a single objective, prior to optimization. Up to 300 parameters must be optimized, satisfying objectives which are often competing. We use multiobjective dose optimization methods where the objectives are expressed in terms of quantities derived from dose-volume histograms or in terms of statistical parameters of dose distributions from a small number of sampling points. For the last approach we compare the optimization results of evolutionary multiobjective algorithms with deterministic optimization methods. The deterministic algorithms are very efficient and produce the best results, but they also have certain limitations. The performance of the multiobjective evolutionary algorithms is improved if a small part of the population is initialized by deterministic algorithms.