Improvement Of Sparrow Search Algorithm Based On Rough Data-Deduction And Application Research

The sparrow search algorithm is a stochastic search algorithm that combines the concept of population intelligence and is designed based on the behavior of sparrow populations foraging and avoiding natural enemies in nature.However,the algorithm faces the problems of poor interpretability,lack of theoretical basis,rapid decline of population diversity during iteration and easy to fall into local optimality.As a continuous population algorithm,the sparrow search algorithm is unable to solve problems in discrete spaces,which leads to its development being limited.To address the above-mentioned defects,this paper proposes appropriate improvement strategies,improves and discretizes the design of the algorithm,and applies the algorithm to the hybrid flow shop scheduling problem.The main work is as follows.(1)A multi-strategy improved sparrow search algorithm based on rough data-deduction is proposed for the problems of poor interpretability,lack of theoretical basis,too rapid decline of population diversity in the iterative process and easy to fall into local optimality of the sparrow search algorithm.Firstly,we combine the idea of low difference sequence for population initialization to enhance the global search ability of the algorithm and guarantee the integrity of the domain of rough data-deduction theory;then we introduce the theory of rough datadeduction,combine fitness and distance to establish the connection between individuals,improve the convergence speed of the algorithm,enhance the ability to jump out of local optimum,improve the deficiency of the sparrow search algorithm in the multi-peak problem,and improve the interpretability of the algorithm;finally For the over-bounded individuals in the iteration,the over-bounded individuals are assigned to the values near the boundary instead of the maximum or minimum values of the boundary to ensure the diversity of the population and improve the convergence speed of the algorithm.Through formal reasoning,the rationality of the improved strategy is demonstrated,and a comparison test with other algorithms illustrates the superiority of the improved algorithm.(2)A discrete sparrow search algorithm is proposed to address the shortcoming that the sparrow search algorithm cannot solve discrete problems.Firstly,the position update formula is abstracted according to the amount of changing individual information and the movement of different individuals of the algorithm;secondly,the design way of the discrete algorithm is studied to compare the search range,exploitation and exploration ability of the heuristic convergence strategy;finally,a new discrete update strategy is designed for the sparrow search algorithm while retaining the original algorithm framework,and the discrete sparrow search algorithm is proposed.(3)Improve the discrete sparrow search algorithm to solve the hybrid flow shop scheduling problem.Firstly,we establish a mathematical model for the hybrid shop-floor scheduling problem and design the encoding and decoding methods;secondly,we integrate the rough datadeduction theory to improve the discrete sparrow search algorithm,expand the search strategy,and improve the solution accuracy;finally,we use the improved discrete sparrow search algorithm to solve the hybrid shop-floor scheduling problem and conduct simulation experiments on three small-scale practical cases and ten large-scale standard test sets.Finally,we demonstrate the feasibility of the improved discrete sparrow search algorithm to solve the hybrid flow shop scheduling problem with higher accuracy compared with other algorithms.