A random benchmark suite and a new reaction strategy in dynamic multiobjective optimization

In the domain of evolutionary computation, more and attention has been paid to dynamic multiobjective optimization. Generally, artificial benchmarks are effective tools for performance evaluation algorithms (DMOEAs). After reviewing existing highlighting their weaknesses, this paper proposes a new benchmark suite promote comprehensive testing algorithms. This proposed eight random instances in which randomness is produced by designed time sequences. Also, introduces challenging but rarely considered characteristics, including diverse features fitness landscape (e.g. deception, multimodality, bias) complex trade-off geometries convexity-concavity mixed geometry disconnected geometry). Empirical studies have shown that poses reasonable challenges DMOEAs terms convergence diversity. Besides, center matching strategy (CMS) suggested track changes these problems, applies history individual information global scope population prediction. Compared with other reaction strategies, CMS demonstrated be very competitive dealing problems.