The advancement in distributed and intelligent computing has facilitated the use of software agents for implementing e-services; most electronic market places offer their customers virtual agents that can do their bidding (i.e., eBay, onSale). E-transactions via shopping agents constitute a promising opportunity in the e-markets (Chen, Vahidov, & Kersten, 2004). It becomes relevant what kind of information and what kinds of bargain policies are used both by agents and by the market place. There are several steps for building e-business: (1) attracting the customer, (2) knowing how they buy, (3) making transactions, (4) perfecting orders, (5) giving effective customer service, (6) offering customers recourse for problems such as breakage or returns, and (7) providing a rapid conclusion such as electronic payment. In the distributed e-market paradigm, these functions are abstracted via agents representing both contractual parts. In recent years, many researchers in intelligent agents’ domain have focused on the design of market architectures for electronic commerce (Fikes, Engelmore, Farquhar, & Pratt, 1995; Schoop & Quix, 2001; Zwass, 1999), and on protocols governing the interaction of rational agents engaged in such transactions (Hogg & Jennings, 1997; Kersten & Lai, 2005). While providing support for direct agent interaction, existing architectures for multiagent virtual markets usually lack explicit facilities for handling negotiation protocols, since they do not provide such protocols as an integrated part of the framework. In this article we will discuss the problem of contract negotiation in e-marketplaces. In the next section, we will present related models commonly used to implement negotiation in e-markets, game theory models, auction models, and contract-net protocols. Then the following section continues with the presentation of a negotiation protocol based on dependency relations. We then present a negotiation strategy based on risk evaluation. The conclusion summarizes the article and paves the further way concerning the truth in the negotiation strategy and the use of temporal aspects on commitments and executions of contracts.
Negotiation Protocols For E-Market
The interaction between agents inside the marketplace is managed by a negotiation protocol. In fact, the negotiation protocol defines a set of public rules that allow agents to set up transaction contracts or co-operation agreements. Previous work and significant achievements are reported on various related fields of research and concrete solutions. Most of the Internet-based market places use auction protocols, especially the English auction.
Hereafter, we present and evaluate some negotiation models either developed in some research works or implemented in some practical systems: game theory, auction models, and contract-net protocols.
Key Terms in this Chapter
Negotiation Strategy: Can be defined as the way in which a given party acts within the negotiation protocol rules in an effort to get the best outcome of the negotiation. It is meanly the process by which the agent evaluates offers and generates counter-offers.
Rational Agents: An agent is individually rational if it accepts only deals that give him or her non-negative utility.
Autonomous Agents: In pure autonomous agents systems, the concern of the designer is with the performance of the individual agent, and the system level performance is left to emerge from the agents’ interactions without taking into consideration the interdependency between the system’s components.
Joint Profit: A combination of the individual profit or individual loss for all agents participating in the contract. If N agents participate in a contract, then JP = Si=1N Pi where Pi is the agent’s Ai profit (positive value) or agent’s Ai loss (negative value).
Cooperative Agents: In pure distributed problem solvers where agents are supposed to be fully cooperative, the only concern of the designer is with the overall performance of the system, and the performance of the individual agent is not important.
Agent’s Mental State: Three main components are in the agent’s mental state: beliefs, desires, and intentions. However, in the market place context we are adding the trust component.
Negotiation Protocol: Set of rules and knowledge structures that provide a means of standardizing the communication between participants in the negotiation process. The negotiation protocol defines how the actors can interact with each other and often includes the way in which offers and messages are constructed and sent to the opponent.
Social Rationality: The principle of social rationality as was stated by Hogg and Jennings (1997, p. 61) is, “If a socially rational agent can perform an action whose joint benefit is greater than its joint loss, then it may select that action.”