Dynamic Service Pricing for Brokers in a Multi-Agent Economy

  • Authors:
  • Affiliations:
  • Venue:
  • ICMAS '00 Proceedings of the Fourth International Conference on MultiAgent Systems (ICMAS-2000)
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

We study the price dynamics in a multi-agent economy consisting of buyers and competing sellers, where each seller has limited information about its competitors' prices. In this economy, buyers use shopbots while the sellers employ automated pricing agents or pricebots. Derivative following (DF) provides a simple, albeit naive, strategy for dynamic pricing in such a scenario. In this work, we refine the DF algorithm and introduce a model optimizer (MO) algorithm that re-estimates the price-profit relationship for a seller at every interval more efficiently. Simulations using the MO pricebots indicate that it outperforms DF even though it has no additional information about the market.