Production Decision Analysis Based on Decision Tree and Bayesian Optimisation

This paper proposes a production decision analysis model based on decision tree and Bayesian optimisation, aiming to optimise the decision-making in the production process of enterprises. Firstly, a spare parts sampling model is established through Bayesian optimisation to determine the optimal sampling sample capacity to minimise the number of inspections and cost. In particular, the binomial distribution of the number of unqualified products is approximated as a normal distribution by using the central limit theorem, and the effect of sampling error on the sample capacity is determined by the hypothesis testing method. Next, this paper constructs a single-objective optimisation model, and uses dynamic programming and decision tree model to minimise the cost loss at different stages. By setting the accessory detection threshold and the finished product detection threshold, the precise decision logic is provided, and the optimal decision at each stage is solved step by step using the set of state vectors. This paper combines Bayesian optimisation with dynamic programming and decision tree model to provide a systematic production decision optimisation scheme, which provides theoretical support for enterprises to make more accurate decisions.