Multi-Stage Production Decisions Based on Monte Carlo and Markov Decision Algorithms

This paper presents a comprehensive multi-stage production decision model aimed at optimizing decision-making processes. Initially, the model is solved using both Markov Decision Processes (MDP) and Genetic Algorithms (GA) to obtain diverse decision outcomes and cost-benefit analyses. A comparative analysis reveals the superiority of the Markov Decision Algorithm in achieving optimal decisions. Furthermore, with defect rates of components, semi-finished, and finished products determined via sampling inspection, a cost-benefit analysis model is developed. A Bayesian updating model is constructed to integrate sampling data with prior information, yielding a more precise estimate of defect rates. Bayesian inference is then employed to recalibrate the model and derive optimal decisions. Ultimately, the Monte Carlo Algorithm and the Markov Decision Algorithm are confirmed as the optimal decision-making approaches for the problem studied.