An agent-based simulation of the effects of consumer behavior on market price dynamics

This paper uses agent-based simulation to explore the impact of consumer behavior on the evolution of market prices in a two-tiered supply chain. In terms of preferences, consumers are sensitive to both product price and retailer quality of service, which measures the retailer's ability to immediately satisfy consumer demand. In this virtual marketplace, consumers are effected by bounded rationality and their limited knowledge of the environment. This is formulated in terms of the visibility they have over the full set of market prices, and retailers' performance. Our model involves a hybrid learning approach. Reinforcement learning is used to model consumer learning about the retailers' reputation for availability. The retailer pricing mechanism is controlled using a genetic algorithm, where the agents compete with each other for higher profits as they continually adapt in an ever-evolving environment. We have examined the impact of the consumer behavior on the evolution equilibrium of market prices and retailer profits. Our simulation results give an insight into the relationship between market price and consumer behavior, and also have potentially interesting applications to real-world marketing strategies.