Heuristics-based intelligent agents for deriving reserve prices in English online auction

Law, Edwin Ban Hock (2009) Heuristics-based intelligent agents for deriving reserve prices in English online auction. Masters thesis, Universiti Malaysia Sabah.


Download (2MB) | Preview


Online auction is popular due to the flexibility and convenience that it offers to e-consumers. Most sellers come and gather in online auction with the aims of obtaining profit. To guarantee the sale of the item with a greater profit, the reserve price for the item must be determined before the item is put up for sale. Here, the reserve price is the minimum price that the seller is willing to sell the item and this price is made hidden throughout the auction. However, setting the price too high may result in no sale whilst setting the price too low may yield in a lower profit. Within the auction context, the process of deriving for the best optimal reserve price is complicated and is not straightforward as the online auction environment is extremely complex. In real auction, most sellers fail to place a strategic reserve price for the auctioned item resulting in a lower profit as simple pricing strategy could have been deployed when auctioning an object. In this work, an autonomous seller agent is developed that generates the item's reserve price using a heuristic decision making framework by exploiting the past bidding history. Here, the reserve price is determined based on several selling constraints that comprise of the number of competitors, the number of bidders, the duration for the auction and also the degree of profit that the seller desires when disposing an item for sale. The detail design and implementation of the agent's strategy is elaborated in this thesis. The strategy will be evaluated across a wide context of diverse and a variety of selling environments. Here, the agent is tested in different circumstances where it was assumed to have complete information about its environment where the agent knows the actual number of sellers and buyers. In real auction however, this information is not known where the number of auctions selling the same item and the number of bidders participating in each auction are entirely not known. Hence, the performance of the seller agent is further investigated in situations where the agents are having incomplete information about the environment. The seller strategy is evaluated under a simulated auction marketplace using an independent private value framework with dynamic participation entry. Result shows that the intelligent agent was able to derive for an optimal reserve price while delivering a high success rate, a high profit, a high selling price, and a high percentage of gain towards the market price. Due to the efficiency of the heuristic algorithm, the agent's strategy could be considered as a new model for single unit object, private reserve price and English protocol.

Item Type: Thesis (Masters)
Uncontrolled Keywords: English auction, heuristic decision making framework, reserve price, seller agent, seller strategy.
Subjects: Q Science > QA Mathematics > QA76 Computer software
Divisions: SCHOOL > School of Engineering and Information Technology
Date Deposited: 10 Jun 2013 07:23
Last Modified: 06 Oct 2017 08:44
URI: http://eprints.ums.edu.my/id/eprint/6220

Actions (login required)

View Item View Item