Developing environmental hedging point policy with variable demand: a machine learning approach

This study evaluates the effect of carbon emission control policies on organizations' production planning and inventory management. Considering the variation of demands, breakdowns, and environmental uncertainties, we consider environmental Hedging Point Policy to control production level in relation to the costs of inventory, backlog, and emission. The effect of Cap-and-Trade, and Command-and-Control environmental policies on product lines’ strategies are evaluated. We aim to develop a production plan through optimization-based simulation and provide a solution for variable demand. Therefore, a simulation-based optimization on multi-objective particle swarm algorithm has been applied (RMSE = 0.82). To acquire practical and managerial implications, through machine learning, the environmental Hedging Point Policy parameters for variable demands are obtained. The results reveal that the Cap-and-Trade policy is more flexible and effective than the Command-and-Control in terms of reducing costs and using environmentally friendly technologies. Our approach offers an effective solution to help decision makers to dynamically plan operations for variable demands, utilize resources, and manage inventories, and increase productivity.