AN AGENT-BASED SIMULATION OF THE EFFECTS OF CONSUMER BEHAVIOR ON MARKET PRICE DYNAMICS

摘要

This paper uses agent-based simulation to explore the im-rnpact of consumer behavior on the evolution of marketrnprices in a two-tiered supply chain. In terms of prefer-rnences, consumers are sensitive to both product price and re-rntailer quality of service, which measures the retailer’s abil-rnity to immediately satisfy consumer demand. In this virtualrnmarketplace, consumers are effected by bounded rational-rnity and their limited knowledge of the environment. This isrnformulated in terms of the visibility they have over the fullrnset of market prices, and retailers’ performance. Our modelrninvolves a hybrid learning approach. Reinforcement learn-rning is used to model consumer learning about the retailers’rnreputation for availability. The retailer pricing mechanismrnis controlled using a genetic algorithm, where the agentsrncompete with each other for higher profits as they contin-rnually adapt in an ever-evolving environment. We have ex-rnamined the impact of the consumer behavior on the evolu-rntion equilibrium of market prices and retailer profits. Ourrnsimulation results give an insight into the relationship be-rntween market price and consumer behavior, and also havernpotentially interesting applications to real-world marketingrnstrategies.