Olivier C.L. Haas

Hybrid optimisation technique for radiotherapy treatment planning

Presents an optimisation strategy, based on a hybrid multi-objective genetic algorithm, to automate the planning of traditional radiotherapy treatment. The optimisation algorithm is formulated around the concept of Pareto optimality to exploit a population based search procedure thereby considering each of the objectives independently. It explores the solutions belonging to the Pareto optimal set prior to concentrating on the best regions of the search space. Once the Pareto optimal set has been determined, a posteriori articulation of the objectives is used to provide the clinicians with a few alternative solutions, all of them non-dominated and therefore mathematically optimal, leaving the final decision to the clinicians.