Genetic Algorithm Based Multi-Objective Optimisation Model in Production Decision Making

In this paper, the production decision optimisation problem is investigated, and a multi-objective optimisation model is solved using Genetic Algorithm (GA). Firstly, for the problem of detecting the defective rate of spare parts, the Bernoulli assumption is used and combined with the binomial distribution, hypothesis testing and the central limit theorem to determine the minimum sampling number to ensure that the defective rate is detected in accordance with the set criteria. Through calculation, a specific table of sampling number and defective critical value is generated. Secondly, the optimisation problem was decomposed into a multi-objective planning model, including maximising the inspection efficiency, minimising the total inspection cost and minimising the loss probability, and the dynamic inspection probability was introduced as a decision variable in the model to replace the traditional 0-1 variable. Through the global search capability of the genetic algorithm, the optimal solutions under six different scenarios and their corresponding inspection schemes and cost analyses were solved. Finally, with the introduction of semi-finished product process optimisation, the key variables were adjusted and the multi-objective optimisation problem was simplified by combining with the ideal point method, and further optimised by the genetic algorithm, and the corresponding cost of the optimal detection scheme was finally calculated to be 3619.17.

I. Introduction

In modern complex systems, how to effectively solve multi-objective optimisation problems by intelligent algorithms has become one of the hot research topics in computer science [1], [2]. Especially in production and quality control scenarios, optimisation problems often involve multiple conflicting objectives, such as improving inspection efficiency, reducing cost, and reducing resource waste. Traditional optimisation methods are difficult to satisfy these diverse needs simultaneously, so advanced computational learning algorithms, such as Genetic Algorithm (GA), are introduced to provide new ideas for solving such problems. Genetic Algorithm, as a global optimisation method based on natural selection and genetic mechanism, has been widely used in the field of multi-objective optimisation due to its powerful parallel computation and searching ability [3], [4].