Research on Optimization of Production Decision Based on Dynamic Planning and Genetic Algorithm

In the production process of enterprises, it is crucial to optimize the decision-making to achieve the maximum profit. In this paper, a mathematical model is constructed by adaptive dynamic programming and genetic algorithm to solve the problem of minimizing the number of inspections and optimizing the production decision within the acceptable defective rate interval [1]–[3]. For minimizing the number of inspections, the model is first built using hypothesis testing and binomial distribution and approximated by normal distribution. Further, the adaptive sequential sampling method (SPRT) is used to derive the minimum number of inspections required to achieve 90% and 95% confidence levels with a set defective rate of 0.1. In production decision optimization, this paper divides the overall process into three stages: inspection of spare parts, finished products and dismantling of defective products. Through the dynamic programming algorithm, each decision is evaluated by backtracking from the third stage to the first stage, and finally the optimal decision and the maximum benefit are derived for the six scenarios [4]. For example, in the case of 200 finished products, the maximum gain is 5560, and the optimal decision is not to inspect parts and finished products, but to dismantle nonconforming products. In addition, for the case of two processes with eight spare parts, the dynamic programming model is applied to solve the problem and the optimal gain for each finished product is 77.91 dollars. To validate the results, a genetic algorithm model is developed and yields a minimized cost of $124.70 with an error rate of only 3%. For the more general multi-process, multi-part problem, the model is reconstructed using genetic algorithms, demonstrating the potential for decision optimization under complex conditions.