A multiple agent-based system for the intelligent demand planning of new products

New product development (NPD) is pivotal in the firm's innovation and organic growth. Despite its strategic importance to the firm's success, it poses many managerial challenges for the effective launch of new products due to the inherent difficulty in new product demand planning. Such difficulty stems from an absence of historical sales data, shortened product life cycles, and a rapid shift in today's consumer behaviours. To deal with those demand planning challenges, this paper aims to propose a multiple agent-based system (ABS) that can overcome the shortcomings of traditional demand forecasting tools and improve forecasting accuracy significantly through the inclusion of meaningful information available from both internal and external data sources. The proposed ABS incorporates causal information obtained from four different types of agents: the coordination agent, the task agent, the data collection agent, and the interface agent. Through a series of simulation experiments, we found that the ABS improved forecasting accuracy over the traditional forecasting methods in demand planning situations where only a limited amount of historical data is available in the early introductory stages of NPD. [Received: 11 August 2022; Accepted: 24 March 2023]