A stochastic approach to full inverse treatment planning for charged-particle therapy (Q785636)

The paper deals with an optimization problem called full inverse treatment planning (full ITP). It deals with finding a radiotherapy treatment plan to tumors that accurately matches the dose prescribed to the tumor while sparing nearby healthy structures. The full ITP is a quite complex optimization problem, full of local minima. For this reason, in the past, research was focused on its simplified version, in which only a reduced set of variables is optimized, compared to the full ITP. The present article tries to fill this gap in the literature by dealing directly with the full ITP. This is achieved by proposing a variation of simulation annealing, which takes into account specific features of full ITP, without altering global convergence properties of the algorithm. Numerical experiments show the effectiveness of the methodology proposed in the article.