Multi Agent Collaborative Search Many-objective Optimization Algorithm

Abstract

MACS (Multi Agent Collaborative Search) is a creative multi-objective optimization algorithm which is effective for handling standard benchmarks and real engineering problems. It has a long improving history and the latest version is Improved archiving and search strategies for Multi Agent Collaborative Search (MACS2.1). But its ability on solving many-objective optimization problems is not good enough. This paper extends the original MACS2.1 to improve its performance to treat those problems and proposes a new multi agent collaborative search many-objective optimization algorithm (named Ma-MACS). Firstly, a more reasonable computing resources allocation approach (utility function) is applied to balance the quality for each individual. Secondly, a neighborhood updating process is embedded to increase evolution speed. Next, more mutation operators and related choosing strategy are applied to enhance the quality of the offspring. Finally, a weight vectors adjusting procedure is introduced to replace the inappropriate vectors. The new algorithm is compared with some state-of-art algorithms and MACS2.1 on the test cases.