Improved Multi-Objective Optimization Algorithm Based On Decomposition And Its Application In Array Antenna Pattern

In practical problems,multiple conflicting targets that need to be optimized at the same time are often encountered,population group algorithm can better solve multi-objective optimization problems.Multi-objective Evolutionary Algorithm Based on Decomposition(MOEA/D)is a prominent representative of this.It decomposes multi-objectives and converts them into multiple single-objective optimization problems.The advantages of search ability,efficient fitness evaluation and good convergence performance have become research hotspots for researchers at home and abroad and are applied in various fields.In this paper,we study the advantages and disadvantages of the MOEA/D algorithm,and try to make use of its advantages through improvement,avoid their disadvantages,and apply them to practical problems.The main tasks include: because the advantage of the MOEA/D algorithm is to simultaneously optimize the sub-problems,reducing the diversity of traditional multi-objective evolutionary algorithms,but the disadvantage is that as the target dimension increases and the number of approximate Pareto solutions obtained by the algorithm increases,resulting in a decrease in the algorithm's ability to search for the optimal solution globally and when the Pareto solution set is more complex,the algorithm has insufficient selection of the optimal solution and easily falls into a local optimum.This paper proposes two different ideas of an improved Multi-objective Evolutionary Algorithm Based on Decomposition(IMOEAD).First,the algorithm expands the ability to search the solution space by introducing a normal distribution crossover operator and using its normal distribution and discrete recombination operations,which significantly improves the quality and diversity of the solution;Furthermore,by introducing the Levi flight to the sub-solutions generated after the reorganization to modify the solutions,the flexibility of searching for the solutions is increased,the accuracy of the solutions is effectively improved,and the performance of the algorithm for searching the global optimal solution is improved;Finally,an optimal solution selection mechanism that solves the Euclidean distance is proposed to filter the specified number of liberated into the archive set,so that the selection of the optimal solution is more uniform and stable.Through the improvement of the above three aspects,the spatial search range of the multi-objective evolution solution is increased,and the stability and accuracy of the solution during the iterative process are improved.Second,the algorithm starts with the generation of weight vectors;In view of the limited number of weight vectors generated by the single lattice point method in the original MOEA/D algorithm,it is not possible to freely generate thecorresponding number according to the problem situation and propose to generate it by means of uniform design.In addition,the polynomial mutation used in the original MOEA/D does not get a better evolutionary individual.By mixing polynomial and non-uniform mutations,the evolutionary solution is more inclined to Pareto.The reason is that during the non-uniform mutation process,the evolution of the optimal solution evolves along the direction of the current optimal solution,and the selection of the current optimal solution still uses the Euclidean distance method in the first improved idea.With the above improvements,the search range of the solution is wider and the solution converges faster.The above two IMOEAD algorithms are tested on the ZDT and DTLZ test function sets,respectively.The experimental results show that the two IMOEAD algorithms are superior to the original algorithm and the other four comparison algorithms in the convergence performance,stability and accuracy of the solution.The two types of IMOEAD were further applied to the pattern optimization based on the array antenna,the experiment used an 8-element linear array,by minimizing the excitation current,the maximum sidelobe level is reduced,the energy of the main lobe is increased,and the directivity is improved to improve the energy transmission efficiency.Co-channel crosstalk,a disadvantage in wireless communication,can be overcome by pointing the zero point to the terminal.Therefore,the application of the two types of IMOEAD in a wireless mobile communication system produces a zero point at a specific position,and the experimental results show that the effect is significant.