Intelligent decision support system for optimizing inventory management under stochastic events

This study examines an optimal pricing policy for managing deteriorating products in a stochastic inventory system. Customer arrival is formulated as a Poisson process with a fixed rate, and customer demand is probabilistically expressed based on the valuation of the product. Typically, customer interest in items decreases over time because of the declining quality of deteriorating products. A dynamic pricing policy is needed to increase retailer profitability under stochastic scenarios. This paper combines the Hamilton–Jacobi–Bellman (HJB) equation with the Particle Swarm Optimization (PSO) approach to realize an optimal solution for the dynamic pricing strategy with specific boundaries using an analytical approach. Furthermore, an artificial intelligence-based High-Order Neural Network (HONN) model is implemented as an intelligent strategy to approximate the complex dynamic behaviors of the stochastic inventory system. The present study aims to fill the research gap by comparing the analytical solution obtained from the hybrid optimization using HJB–PSO and the intelligent strategy obtained from the HONN model. The proposed approaches bring greater accuracy and efficiency to inventory management tools. The performance difference between the proposed algorithms indicates that the error does not exceed the desired limit (≤\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le$$\end{document} 1%), indicating that the analytical and numerical solutions found can be used interchangeably in practice. Furthermore, all validation tests ensure that the proposed decision support system can aid policymakers in managing stochastic inventory systems with minimal effort and adopting strategies to ensure resilience and customer satisfaction.