Constrained Multiobjective Optimization Algorithm Based on Immune System Model
提出一种模拟生物免疫系统的优化算法，用于解决带多模态非线性约束的多目标优化问题，通过将种群分为可行非支配和不可行/支配两类分别处理，并在基准测试中表现优于多种现有算法。
An immune optimization algorithm, based on the model of biological immune system, is proposed to solve multiobjective optimization problems with multimodal nonlinear constraints. First, the initial population is divided into feasible nondominated population and infeasible/dominated population. The feasible nondominated individuals focus on exploring the nondominated front through clone and hypermutation based on a proposed affinity design approach, while the infeasible/dominated individuals are exploited and improved via the simulated binary crossover and polynomial mutation operations. And then, to accelerate the convergence of the proposed algorithm, a transformation technique is applied to the combined population of the above two offspring populations. Finally, a crowded-comparison strategy is used to create the next generation population. In numerical experiments, a series of benchmark constrained multiobjective optimization problems are considered to evaluate the performance of the proposed algorithm and it is also compared to several state-of-art algorithms in terms of the inverted generational distance and hypervolume indicators. The results indicate that the new method achieves competitive performance and even statistically significant better results than previous algorithms do on most of the benchmark suite.