Impact of Trust on Agent-Based Simulation for Supply Chains Andr´e Jalbut [0000-0003-1545-6345] and Jaime Sim˜ao Sichman [0000-0001-8924-9643] Laborat´orio de T´ecnicas Inteligentes (LTI) Escola Polit´ecnica (EP) Universidade de S˜ao Paulo (USP) {andre.jalbut,jaime.sichman}usp.br Abstract. Companies in supply chains have a goal to optimize their productivity, and hence their profits. One way to study the behavior of these chains is to simulate them using a multi-agent approach. In this work, we propose an extension of a model used in the literature, called the Beer Game, by adding multiple agents in each of its levels to evaluate both the local and global performance of the suppliers. We use different agent profiles, based either on trust or on price. By enabling clients to ask supplier sugestions from their peers, we measure how these peers’ suggestions and lies affect working capital and trust measures for different agent profiles. 1 Introduction The main goal of every business is profit [6]. In the context of companies inter- acting in a supply chain (SC), partnerships based on trust can be more profitable than those based on supply and demand mechanisms. This statement is based on the observation that the greater the trust of a consumer in his suppliers, the greater the responsiveness of these, and therefore the greater the gain for SC [8]. Supply chains are defined as the set of organizations, activities, information and resources involved in the movement of a product or service from suppliers to consumers [14]. The interest in studying the management of these chains, supply chain management (SCM), has been increasing in order to obtain competitive advantages for the market through improvements in its processes [3]. Figure 1 [11] illustrates an example of SC.