Pruning algorithm for Multi-objective optimization

Because of non-existence of an ideal single solution in Multi-objective optimization frameworks, the set of optimal solutions is required to be well spread and uniformly covering wide area of Pareto front. The decision maker (DM) still work hard to compromise the trade-offs solutions based on his/her preferences. In this paper, we proposed a pruning algorithm that can filter out undesired solutions and provides more robust trade-offs solutions to the DM. Our algorithm is called adaptive angle based pruning algorithm with bias intensity tuning (ADA). The pruning rationale is increasing the dominated area for the purpose of removing solutions that only marginally improves in some objectives while being significantly worse in other objectives. The extra angles are expanded from the regular dominated area. The bias intensity parameter (W) is introduced in order to approximate the portions of desirable solutions based on DM's opinions. We chose several benchmark problems with different difficulties including two and three objectives problems. The experimental result has shown that our pruning algorithm provides robust sub-set of Pareto-optimal solutions on several benchmark problems. The pruned Pareto-optimal solutions distributed and covered multiple regions instead of single region of Pareto front. In addition, it's clearly shown in bi-objective problems that the pruned Pareto-optimal solutions are located at knee regions of the Pareto front.