A dynamic evolutionary multi-objective optimization algorithm based on decomposition and adaptive diversity introduction

In addition to the need for satisfying several objectives, many problems in the real-world are also time-varying. These dynamic problems need the optimization algorithm to constantly find the time-varying optimal solutions. This paper proposes a dynamic multi-objective evolutionary algorithm based on decomposition and adaptive diversity introduction (denoted by dMOEAD-DI). Specifically, the dMOEAD-DI decomposes a dynamic multi-objective optimization problem into several dynamic single objective optimization subproblems and solves them at the same time. Also, an adaptive diversity introduction (ADI) method is adopted to respond to the environment change, and the scale of diversity introduction can be automatically calculated by the real degree of dynamic change, so as to enable the population adapt to the new environment quickly. The effectiveness of the dMOEAD-DI algorithm is validated against other two well-known algorithms. Simulation results on four benchmark problems demonstrate that the dMOEAD-DI has better convergence and distribution of non-dominated solutions obtained around the changing Pareto optimal front.