MoCAAS: auction agent system using a collaborative mobile agent in electronic commerce

https://doi.org/10.1016/S0957-4174(02)00141-0Get rights and content

Abstract

To get the items that a buyer wants in an Internet auction, he must search for the items through several auction sites. When the bidding starts, the buyer needs to connect to these auction sites frequently so that he can monitor the bid states and re-bid. A reserve-price auction reduces the number of connections, but this limits the user's bidding strategy. Another problem is equity between the buyer and the seller. Both the buyer and the seller should profit within proper limits. In this paper, we propose an auction agent system using a collaborative mobile agent and a brokering mechanism called MoCAAS (Mobile collaborative auction agent system), which mediates between the buyer and the seller and executes bidding asynchronously and autonomously. This reduces the network load more than with other auction-agents, offers more intelligent bidding, and increases the clear ratio.

Introduction

Though ECs technology has grown steadily, it is still difficult to implement the negotiation between a buyer and a seller online. The Internet auction has been widely expanded as an alternative solution (Lee, Choi, Kim, & Lee, 1999). However, to search for items, monitor bid states, and re-bid, users need to connect to the auction site frequently. Also, first-time buyers do not know bidding strategies nor an item's value; therefore a competitor may cheat them, and they may lose the chance to buy items at a cheaper price. As a result, they may not buy the items that they want, or they may pay too much.

When the above problems are solved, the Internet auction will become a generalized EC market. Thus, in this paper, we propose an auction agent system called MoCAAS (Mobile collaborative auction agent system), which mediates between the buyer and the seller and executes bidding autonomously for the buyer. When a buyer submits a reserve-price and the identity of an item, the agent searches for the item among registered auctions. It then recommends auctions to the buyer and informs the buyer of the expected price for the item. When the buyer selects the best among the recommended auctions, the agent executes bidding for the buyer.

This section of our paper presents an overview of the MoCAAS system and its benefits. Section 2 presents an overview of the auction, the auction agent, and the collaborative mobile agent. Section 3 presents the architecture and workflow of MoCAAS. Section 4 presents the bidding processes of MoCAAS. Section 5 presents the brokering algorithm of MoCAAS. Section 6 reports experimental evaluation results. Finally, Section 7 presents a brief summary and future considerations.

Section snippets

The English auction

The ‘auction’ is the buying and selling of property through public bidding. The ‘English auction’ is the most common and simplest type of auction.1 Sotheby's and Christie's use this method for auctioning fine art. This is the method used at most Internet auction sites. In the English auction, the auction house will take bids in ascending order, and a bidder must bid more than the ‘going price’. The highest bidder receives the item and pays for

Architecture and workflow

MoCAAS consists of five main components:

  • a buyer-agent,

  • a broker-agent,

  • a bid-agent,

  • an auctioneer-agent.

The architecture of a MoCAAS is shown in Fig. 1. The buyer-agent offers a buyer interfaces for querying the broker-agent, specifying the bid-agents, controlling the bid-agents. An interface for querying the broker-agent shows a recommended auctioneer list and the expected price of the item. An interface for specifying agents sends a bid-agent creator the information for creating the bid-agent.

Brokering algorithm

CDA matches multiple buyers and sellers. Therefore, the MoCAAS's brokering algorithm is similar to the CDA's. AuctionBot implements CDA with Mth-price and (M+1)st-price rules (Wurman, Wellman, & Wash, 1998). In a set of L single unit bids, M is ‘sell offers’ and the remaining N=LM are ‘buy offers’. The Mth-price auction clearing rule sets the price at the Mth highest among L. Similarly, the (M+1)st price rule chooses the price of the (M+1)st bid. Mth price is undefined if there are no sellers,

Bidding process

The dispatched bid-agent executes the bidding autonomously by exchanging messages between the bid-agent and the auctioneer-agent. (Fig. 4) First, when the bid-agent arrives at an auction host, the bid-agent registers itself with an auctioneer-agent in the auction host. (1→2) If the bid-agent is valid, the auctioneer-agent accepts the bid-agent and updates the information of accepted bidders. (2→3) When the bidder submits a bid, (3→4) the auctioneer-agent evaluates the bid. If the bid is valid,

Experimental evaluation

We selected an arbitrary item and compared MoCAAS's performance with the performance of other existing auction agents. First, we created a scenario.

  • Every 30 min for 6 h, a user monitors the auction that he (or she) is participating.

  • He re-bids four times during that time.

A stationary agent and the MoCAAS's mobile agent executed the above behaviors. After this was completed, the network load was measured. In the case of using the mobile agent, after dispatching the agent, network load didn't

Conclusions

We have designed and implemented a system whereby a mobile agent substitutes for a buyer efficiently. A mobile agent mechanism reduces user operations and network load. The broker-agent serves to implement the coordination of bids across multiple auction sites. This agent uses a modified Mth-price, (M+1)st-price algorithm so that it increases the clear ratio. In the wireless Internet environment, communication costs rise in proportion to a rise in network load. The effect is greater than in the

Acknowledgements

This work was supported by a grant No. 2000-1-30300-016-3 from the Korea Science and Engineering Foundation.

References (6)

  • J.K Lee et al.

    Principles of electronic commerce

    Bupyoungsa

    (1999)
  • T Sandholm et al.

    Nomad: mobile-agent system for an internet-based auction house

    IEEE Internet Computing

    (2000)
  • E Steinmetz et al.

    Bid evaluation and selection in the MAGNET automated contracting system

    First International Workshop on Agent Mediated Electronic Trading AMET-98

    (1998)
There are more references available in the full text version of this article.

Cited by (0)

View full text