Purchase intention-based agent for customer behaviours

Simulating human activities remains a challenging problem because the decision-making mechanisms underlying these activities are difficult to reproduce and mimic. In this article, we are interested in the simulation of in-store shoppers whose activities are generally divided into two parts: a walking activity and a purchase activity. Since the act of buying is more complex than simply following a shopping list, we propose here to model the attraction relationships that can exist between a product and a customer. This attraction model is used to build a multi-agent simulation whose realism is evaluated through various experiments. (C) 2020 Elsevier Inc. All rights reserved.