Optimization of the Multi-Threaded Interval Algorithm for the Pareto-Set Computation

Abstract

Previous investigations of the authors surveyed the possibility of applying interval methods to seek the Paretofront of a multicriterial nonlinear problem. An efficient algorithm has been proposed and its implementation in a multicore environment has been done and tested. This paper has two goals. First one is to tune the developed algorithm to increase the speedup of the multi-threaded variant. The second one is to extend the algorithm to compute not only the Paretofront (in the criteria space), but also the Pareto-set (in the decision space). Numerical results for suitable test problems are presented.