Collaborative optimization by shared objective function data

This poster presents a collaborative algorithmic framework for solving a set of different, yet similar, single-objective optimization problems. Common for these problems is that the evaluation of their objectives consists of two parts: The first part involves a computationally heavy task, like numerical simulation, while the second part further evaluates the objective by performing additional, significantly less computationally-intensive calculations. The idea behind the collaborative framework is (i) to solve all problems in the set simultaneously and (ii) at each iteration, to perform a synchronous “collaborative” operation. This special operation entails sharing the outcome of the heavy part between all search processes. The goal is to improve the performance of each individual process by taking advantage of the already-computed heavy part of solution candidates from other searches. A problem set is presented. With respect to solution quality, consistency, and convergence speed, we observe that our collaborative algorithms perform better than traditional optimization techniques. In particular, information sharing is seen as essential during early stages of optimization. Though the collaborative algorithms require additional computing time, this cost is negligible if the heavy part is much more computationally expensive than the light part.