Multi-objective Evolutionary Algorithms Based On Decomposition And Uniform Design And Their Applications

In real life,multi-objective optimization is becoming more and more common,multi-objective evolutionary algorithms are the most common and effective method to solve multi-objective optimization problems.The objective of the multi-objective optimization algorithm is to quickly obtain a set of solutions with good convergence and diversity.In this thesis,a multi-objective evolutionary algorithm based on decomposition and uniform design is designed to improve the efficiency of the algorithm.The algorithm is also applied to solve the syngas production problem.A good search strategy helps a lot to get high-quality solutions.In this thesis,a multi-search strategy(including three strategies)is used to balance local search and global search and improve the search efficiency.The first search strategy selects the convergent and sparse solutions and the adjacent solutions as the parent solutions to generate new solutions.The second search strategy selects the dominant solutions and the adjacent non-dominant solutions as the parent solutions to generate new solutions.The third search strategy selects non-dominant solutions as the parent solutions to generate new solutions.Uniform design samples uniformly in the sample space.In this thesis,a crossover operator based on uniform design is designed to sample in irregular sampling space,so as to improve the search efficiency.At the same time,in order to satisfy the convergence and diversity of solutions,we adopt an update strategy based on decomposition.Finally,we apply this algorithm to the syngas production problem,and the experimental result shows that the performance of the algorithm is better than most algorithms.Through the test of common test functions and performance comparison with MOEA/D,MOEA/DD,NSGAII and MOEA/D-DRA,concluded that MOEA/DU is better than other algorithms in convergence and diversity.