Multi-Agent Automated Intelligent Shopping System (MAISS)

Agent Technology continues to be one of the most cherished research areas nowadays. In particular, the multi-agent field related research topics and applications have increased dramatically. This is due to the success of intelligent agents which can act on the behalf of a human being. The human involvement will be for monitoring and assigning tasks to agents, which in turn will do other specific operations in a very fast and accurate manner.The use of Agent Technology to support Ecommerce operations, especially in automating the buying and selling process is so promising and is worth of success. Electronic commerce grows rapidly over the Internet. The information on internet becomes more dynamic and heterogeneous. Thus, software agents are required to provide a strong working structure as they help build powerful distributed systems.This thesis studies, analyses and addresses in detail the possible scenarios which are needed for analyzing, designing and implementing a multi-agent automated intelligent shopping system which can handle ecommerce operations. Agents in interaction must have some coordination and communication mechanisms as it will be defined by the Reinforcement learning algorithms and game theory techniques. By coordinating, agents will perform better to achieve a common goal. Agents will therefore have some strategies by which an action taken by each agent is the best action for it, and a joint action by best actions is the best strategy for them to achieve their goal with success.The proposed multi-agent automated intelligent shopping system (MAISS) is a distributed system, where human users (buyers and sellers) are able to delegate their tasks to agents, which will then do the shopping job on their behalf and present them the results. Buyer users (customers) and the seller users (suppliers) can be organizations, companies or individuals.Software agents are capable of selling and buying as long as they are still alive. They can stay longer at the market price than a human could do. Buyer agents are intelligent enough to search for seller agents, negotiate with them and buy after they reach an agreement. Seller agents can advertise their products on the market place. If buyer agents are available, seller agents will communicate with them, negotiate about the price and then after reaching an agreement are reached, the seller agent will sell a product to the buyer offering a good price.The approach of multi-agent automated intelligent shopping systems contributes to finding optimal solutions to enhance and facilitate e-commerce transactions (including automated price negotiations).