Artificial intelligence-based inventory management: a Monte Carlo tree search approach

Abstract

The coordination of order policies constitutes a great challenge in supply chain inventory management as various stochastic factors increase its complexity. Therefore, analytical approaches to determine a policy that minimises overall inventory costs are only suitable to a limited extent. In contrast, we adopt a heuristic approach, from the domain of artificial intelligence (AI), namely, Monte Carlo tree search (MCTS). To the best of our knowledge, MCTS has neither been applied to supply chain inventory management before nor is it yet widely disseminated in other branches of operations research. We develop an offline model as well as an online model which bases decisions on real-time data. For demonstration purposes, we consider a supply chain structure similar to the classical beer game with four actors and both stochastic demand and lead times. We demonstrate that both the offline and the online MCTS models perform better than other previously adopted AI-based approaches. Furthermore, we provide evidence that a dynamic order policy determined by MCTS eliminates the bullwhip effect.

Inventory Optimization Model Design with Machine Learning Approach in Feed Mill Company

“…And for continuous learning, it is necessary to enter data cyclically [21]. Then, processing takes place, a step in which the algorithms organize the data.…”

“…The unsupervised system can verify grouped data and recognize patterns in them [9]. The system by reinforcement occurs with the entry of game-like rules, hence the system learns by trial and error until it specializes in the task, it can also be considered a technique of positive reinforcement [6] and also have transfer learning, which corresponds to the reuse of an algorithm that was used for a task that can be reused in another without the need for development from the beginning [12].…”

“…Machine learning as an necessarily portion of cleverly frameworks is broadly connected in different zones of fabricating such as optimization, control and quality. The objective of the machine learning method is the discovery of designs that depict connections and structure between information components [6].…”

“…[20]. The fast adjustment to the environment of machine learning calculations for the procurement of future information of a framework bolster supervisors in making choices in fabricating forms with the point of progressing the execution of fabricating forms [21]. Design discovery and acknowledgment of regularities of existing information connections are potential highlights of machine learning.…”

“…Later progresses within the field of machine learning (ML)a category of manufactured insights methodscan contribute to decreasing vulnerability in deals by anticipating them from accessible information and by finding complex connections between different components and past exchanges. Through ML, it is possible to model how these elements influence sales and make daily forecasts in the case of small companies, bringing greater clarity to the planning and operation of the business, as well as improving the definition of commercial policies, strategies and tactics [6].…”