A Distributed Agent-based Approach for Robust Optimization

Structural design and optimization in engineering address increasingly non-standard optimization problems (NSP). These problems are characterized by complex topology conditions of the optimization space (w.r.t. nonlinearity, multimodality, discontinuity etc.). By that, NSP can only be solved by means of computer simulations. Hereby, the corresponding numerical approaches applied often tend to be noisy. Examples for NSP occur in robust optimization, where the solution has to be robust with respect to e.g. implementation errors, production tolerances, uncertain environment conditions etc. However, a generally applicable strategy for solving such problem types always equally efficient is not available. To improve the situation, in this paper a distributed agent-based optimization approach for solving NSPs is introduced. The approach elaborated consists of a network of cooperating but also competing strategy agents that wrap various optimization methods (e.g. SQP, DE, ES, PSO etc.) using different search characteristics. In particular, the strategy agents contain an expert system modeling their specific behavior in an optimization environment by means of rules and facts on a highly abstract level. For managing the complexity of NSPs using MAS efficiently, a simulation and experimentation platform has been developed. Serving as a computational steering tool, it applies MAS technology and accesses a network of various optimization methods. As a consequence, an elegant interactive steering, a tailor-made modeling and a powerful visualization of structural optimization processes can be carried out. To demonstrate the far reaching applicability of the proposed approach, numerical examples are discussed, including function and robust optimization problems.