Research On The Application Of An Improved Particle Swarm Optimization Algorithm To Ceramic Formula Problem

Ceramics are one of China’s excellent traditional cultures.As art pieces in ancient times,craftsmen conducted numerous experiments to improve the aesthetics of ceramic products and summarized many ceramic formulas.In modern industry,ceramic products are widely used as excellent non-metal functional materials in modern industries such as architecture,aerospace,automobile,and military industries.Unlike ancient times where ceramic raw material ratio was used as the ceramic formula,modern ceramic formulas are composed of chemical compositions.Solving the ceramic formula problem is a typical optimization problem,which involves selecting the ingredient ratio based on the raw materials and the target formula to minimize the chemical composition error.In the past,mathematical methods such as linear programming,gradient descent,and Newton’s method were commonly used to solve optimization problems.With the emergence of evolutionary computation,such as genetic algorithm and particle swarm optimization algorithm,scholars favor the latter due to its strong optimization capability and simple operation.This article summarizes the impact of three important parameters in particle swarm optimization algorithm on algorithm performance and proposes the following improvements for the problems of particle swarm optimization algorithm being prone to local extreme values and slow convergence speed in the later stage of the algorithm:1.The research status of particle swarm optimization at home and abroad,the generation of ceramic formula problem and the significance of using intelligent optimization algorithm to solve ceramic formula problem are expounded.2.Adaptive inertia weight changes:based on the historical experience and flight status of particle flight,a new adaptive inertia weight strategy is designed by introducing the global range value to adaptively change the inertia weight of each particle in the particle population.When the fitness value of a particle is closer to the global optimal value,it can be judged that this particle needs to explore its neighboring area more.When the fitness value of a particle is closer to the global range value,it can be judged that this particle needs to expand its search range more.3.Non-linear asynchronous learning factors:using larger₁and smaller₂in the early stage of the algorithm can make the particles more focused on self-awareness and try to expand their flight range in the solution space to increase population diversity.In the later stage of the algorithm,using smaller₁and larger₂can speed up the flight speed of the particles towards the global optimal position,effectively balancing the global and local search performance of the particle population.4.Applying the Improved particle swarm optimization algorithm to the ceramic formula problem and converting it from a multi-objective problem to a single-objective problem using linear weighting,combined with Critic method for different chemical composition components weighting。multiple sets of simulation experiments on ceramic formulas have shown that the Improved particle swarm optimization algorithm has significant advantages over other tested algorithms in terms of convergence speed and formula accuracy,and can be effectively applied to ceramic formula design problems.