Research On Sparrow Search Algorithm And Its Application In Boiler Combustion Modeling Optimization

Inspired by the group behavior of sparrow in nature,researchers proposed a new meta heuristic algorithm sparrow search algorithm(SSA).The algorithm is easy to understand,with few parameters and strong local search ability.However,in the face of complex highdimensional function problems,similar to the common heuristic algorithm,there are still "premature" phenomenon and low convergence accuracy.This paper analyzes the shortcomings of SSA algorithm,adds a variety of strategies and mechanisms,improves the optimization range and convergence ability of the algorithm,reduces the probability of falling into local optimum.At the same time,the sparrow search algorithm is extended,and the multi-objective sparrow search algorithm is proposed to solve the multi-objective optimization problem.Finally,the algorithm is applied to the modeling and optimization of boiler combustion system in power plant.The specific research contents of this paper are as follows:By analyzing the mathematical model of sparrow search algorithm and its optimization formula,a Multi Strategy hybrid sparrow search algorithm is proposed.By introducing adaptive parameter control strategy,social learning strategy and spiral search strategy,the convergence speed and the ability to jump out of local optimum of the algorithm are improved.The results are compared with several representative optimization algorithms in 23 test functions and 2engineering optimization problems,and the Wilcoxon statistical test is carried out.Simulation experiment show that hybrid sparrow search algorithm can effectively improve the convergence accuracy and speed compared with other heuristic algorithms and sparrow search algorithm.Aiming at the multi-objective optimization problem,a multi-objective sparrow search algorithm is proposed based on the sparrow search algorithm and the concept of multi-objective.Pareto domination and external archiving mechanism are introduced to improve the diversity of Pareto solution set distribution.Through the contribution of particles to the external archive,the proportion of discoverers is adjusted to improve the convergence speed of the algorithm.In order to verify the effectiveness of the multi-objective sparrow search algorithm,some functions of ZDT series and DTLZ series are selected to compare with the existing multiobjective algorithm.The simulation results show that the MOSSA algorithm has achieved better results than other algorithm in solving multi-objective problems,and can obtain better Pareto solution set.The above two algorithms are applied to the multi-objective optimization of NOx concentration and boiler efficiency in boiler combustion system.Before that,particle swarm optimization algorithm,sparrow search algorithm,gray wolf algorithm and Multi Strategy hybrid sparrow search algorithm were used to optimize the kernel parameters and regularization parameters of LSSVM respectively,and the regression prediction model of NOx emission concentration and boiler efficiency was established.By comparing the accuracy and fitting degree of the regression model,the least squares regression model optimized by hybrid sparrow search algorithm was selected Model is the basis of the model.By setting the termination condition of the multi-objective sparrow search algorithm,the regression model is optimized,and the final multi-objective optimization result is obtained,which proves the effectiveness of the method.