C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....quantity and or quality of the purchased good (cf. 16] The distinguishing aspect of second degree price discrimination is that customers can self select the best purchase. Traditionally, customers are o#ered a menu of options where a tari# assigns a price to the option in the menu. The work of [3, 9] discusses algorithms which given a particular tari # structure learn the best tari#s on line. They conclude that (especially in a dynamic environment) complex tari#s are generally not the most profitable strategy. The distinguishing aspect of the developed system is that instead of having ....

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning complex price schedules. In Proceedings of the AMC Conference on Electronic Commerece, pages 31--41. the Association for Computing Machinery (AMC), ACM Press, 1999.

....(reinforcement learning) Their specific algorithm for the derivative following strategy was adapted for f mite markets and will be analyzed in the Learning Curve Simulator. Their work has provided a strong background for this investigation of successful strategy development. Brooks et al. [19] also performed analysis of pricing agents in a simulated market environment and discussed the Wade offs between exploitation and exploration pricing techniques on the part of the seller. They conclude that when a pricing agent is interested in maximizing revenue over a longer period than the ....

C. H. Brooks, S. Fay, R. Das, J. MacKie-Mason, J. Kephart and E. Durfee. 'Automated Strategy Searches in an Electronic Goods Market: Learning and Complex Price Schedules.' Proceedings of the ACM Conference on Electronic Commerce (EC '99), Denver, CO, 1999.

....before displaying a price when requested by a customer [GKT99] This process should be fast enough to avoid elongating too much the response time perceived by the customer. So, a site may not wait for all the answers before making a decision. A discussion on pricing strategies can be found in [BFDMKD99, Odlyzko99]. Pricing is also an important 4 element when an e business site sells bundles of items as opposed to single items [Parkes99] Privacy is another important consideration. Few sites post privacy statements. These statements tend to be long and complex and are seldom read or understood by users. A ....

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee, Automated Strategy Searches in an Electronic Goods Market: Learning and Complex Price Schedules, Proc.

No context found.

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.

No context found.

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.

....then show experimental results that suggest that congregations can improve the overall utility of the agents in a multiagent system. We will conclude with future directions for studying the formation and dynamics of congregations. 2. CONGREGATING IN INFORMATION ECONOMIES Our past research (e.g. [4, 5]) has focused on how producers in an information economy can locate consumer niches and e#ciently learn the preferences of consumers in these niches. It is this first problem, that of locating a niche, that is most relevant to congregating. In this problem, two or more producers are each ....

....of the experiment, each producer randomly chooses a category of good to sell. Each consumer also has a favorite category (drawn from a uniform distribution) and preferences over non favorite goods indicated by equation 1. Reservation values for a consumer s most preferred good are drawn from U[5, 10]. The number of marketplaces is fixed at the beginning of the experiment; it is assumed that each agent knows m, the number of marketplaces, and each agent has agreed to share equally in the costs of computation in determining an allocation within its market. We assume that computation has a cost ....

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.

....must assess the tradeoff between profitability and complexity for each, compute their expected economic performance, and select the best. In prior work we have studied the profitability complexity tradeoff for a variety of information good pricing schedules and compared their economic performance [5]. We did so for a stationary environment: a single producer offered price schedules to a static population of consumers. Thus, given enough time, the producer could eventually find the maximal point on the profit surface induced by any of the price schedules. However, the best price schedule was ....

....the curve to be fit more closely, but increase the problem s complexity. Table 1 describes the six schedules we have examined and their expected surplus as a function of the bundle size and the population size . In the limit of , an exact analysis of the schedules is possible [5]. These values, scaled by dividing by , are given in column 3 of Table 1. Column 4 of Table 1 gives the peak profits for finite consumer populations, averaged over 10,000 landscapes, each of which was generated by choosing 100 random consumers from the given distribution. The peaks were ....

[Article contains additional citation context not shown here]

C. H. Brooks, S. Fay, R. Das, J. K. MacKie-Mason, J. O. Kephart, and E. H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.

....are rational utility maximizers. That is, they know their preferences and, on every iteration, will act so as to maximize their expected utility. We assume consumer preferences follow a simple two parameter model originally introduced by Chuang and Sirbu [5] and described in our previous work [4, 13]. This model has the advantage of being analytically tractable while still providing significant nonlinearities in consumer demand when consumers are heterogeneous. The model consists of two parameters: which indicates a consumer s value for its most preferred article, and , which indicates ....

....strategy for dealing with consumer heterogeneity [5, 7] If an article is valued by any consumer, including it in a bundle increases the bundle s aggregate value. The calculation of optimal profit with perfect information for each schedule is described in more detail in our previous work [4]; we simply summarize those results here to provide a benchmark for the best possible profit that could be obtained from a schedule once all information is learned. Table 1 shows that as the number of pricing parameters increases, the potential profit a schedule can extract is increased. We refer ....

Christopher H. Brooks, Scott Fay, Rajarshi Das, Jeffrey K. MacKie-Mason, Jeffrey O. Kephart, and Edmund H. Durfee. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, pages 31--40, 1999.

....Wide Web and the use of automated agents to buy and sell goods on behalf of human users has led to the study of agent based information economies. Our previous work in this area has examined the problem of how sellers in an information economy can learn the preferences of a consumer population (Brooks et al. 1999; Brooks, Durfee, Das 2000) In particular, we study problems where the consumer population is nonstationary. We assume that each information goods producer is interested in maximizing its aggregate pro t. Since a producer will only have a limited number of iterations to interact with a consumer ....

Brooks, C. H.; Fay, S.; Das, R.; MacKie-Mason, J. K.; Kephart, J. O.; and Durfee, E. H. 1999. Automated strategy searches in an electronic goods market: Learning and complex price schedules. In Proceedings of ACM EC-99, 31-40.

No context found.

Brooks, C. H., S. Fay, R. Das, J. MacKie-Mason, J. Kephart, and E. Durfee, "Automated Strategy Searches in an Electronic Goods Market: Learning and Complex Price Schedules," Proceedings of the ACM Conference on Electronic Commerce (EC '99), Denver, CO (November, 1999).

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC