Research On Many-Objective Optimization Algorithm Based On Angle Decomposition

Many objective optimization problems are widely used in many engineering fields.However,it has become a major problem in the field of evolution due to the frontier sensitivity of many objective optimization problems.At present,the mainstream algorithms for solving such problems are dimension reduction method,loose dominance method and decomposition method.The former two have the problems of high computational complexity and slow solution efficiency.At the same time,many objective evolutionary algorithms still have defects such as poor convergence.Aiming at the above problems,a new many-objective optimization algorithm called NSGA-?-WA algorithm proposed and it is applied to multi-service quality optimization problems of wireless sensors network.The main research contents of this paper are as follows.The evolution strategy from the many-objective optimization algorithm and the many-objective algorithm frame are studied.Among them,the evolutionary strategy is the key to improve the convergence speed and convergence precision of many-objective optimization algorithms.Therefore,the evolution strategy is improved can effectively improve the performance of the algorithm.It is worth noting that there is no optimal individual in high-dimensional space.A new differential evolution strategy is proposed generates new individuals.It enhances the convergence by adjusting the algorithm's mutating phase and improving the probability selection mechanism for individuals.It also enhances the exploration ability of weight vectors in each subspace.In the many-objective framework,the sensitivity of the individual to the frontier surface increases as the dimension of the objective space increases,and the complexity of the solution set plane in practical applications make it difficult to achieve good distribution.Therefore,the adaptive weight vector adjustment strategy is proposed.By decomposing the target space into several subspaces,the weight vectors are sparse or densely adjusted according to the density of each subspace,ensuring that the weight vectors are evenly distributed on the objective frontier surface.Thereby the distribution of the solved set is achieved.The improved evolutionary strategy is effectively combined with the weight vector adjustment framework.The algorithm is tested on the DTLZ standard test function set and the 3 to 15 target optimization problem of the WFG instance,and compared with the four algorithms with better results.Experimental data shows that the NSGA-?-WA algorithm proposed in this paper is superior to the original algorithm and three other algorithms with excellent performance in convergence and distribution performance.NSGA-?-WA algorithm is applied to multi-service quality optimization problems of heterogeneous wireless multimedia sensor networks oriented to the three-dimensional directed sensing model to verify the effectiveness of the algorithm.The network number,network life cycle,coverage rate and energy consumption are four objectives.The indicators are optimized simultaneously.A heterogeneous wireless network composed of ordinary wireless multimedia sensor nodes and high-energy nodes is created,and the TOPSIS method is used in each round of communication to select an optimal routing connection scheme suitable for the current network topology.The simulation results show that compared with the current performance of NSGA-? and MOEA/D algorithms,the proposed algorithm can balance each objective to achieve better results under different perceptual radii and different perception perspectives.The experimental results verify the effectiveness of the proposed algorithm.