Modeling and implementation of an artificial electricity market using agent-based technology

https://doi.org/10.1016/j.physa.2005.02.068Get rights and content

Abstract

This paper focuses on modeling power exchanges in a multi-agent interacting framework with reduced behavioral assumptions. A model of the day ahead market session of OMEL (the Spanish Power Exchange) is proposed using real demand data with simulated seller strategies. The number of sellers is defined at the first stage and the quantity of goods is distributed over the population of agents according to several initial distributions. A Clearing-house mechanism matches the cumulative demand and supply curves in order to determine the market-clearing price. The resulting price time-series are statistically tested to verify the validity of the model. Results show the main properties of real market and assess the validity of the proposed model.

Introduction

European directive 96/92/CE has demanded a progressive deregulation of production and dealing of electricity in Europe. In many European countries, the state-owned monopolistic production of electric power has been abandoned and electricity is now a commodity exchanged between private producers and consumers. The sale of electric power is made by bilateral dealing or through organized markets. Electric power exchanges are conceived on the basis of different market needs and scopes. Two main categories of power exchanges (PEs) can be identified, i.e., “physical” and “financial” PEs.

Physical PEs are markets which aggregate the effective supply and demand of electricity. Day ahead market (DAM) is the first session of this complex mechanism. It collects and orders all the offers, determining the market price by matching the cumulative demand and supply curves for every hour of the day after according to a merit order rule. Financial PEs are not requested to provide any indication for the unit commitment. They provide a market with standardized and transparent rules where operators can exchange baseload or peakload contracts (standard products). Furthermore, the spot price market allows to define an underlying asset for derivative markets and hedging tools.

Examples of physical PEs are the Spanish market (OMEL) and the Italian market (IPEX); conversely, financial PEs are the German market (EEX), the French market (PowerNext), the new English market (UKPK) and the Netherland one (APX).

This work proposes an heterogeneous agent-based model for a DAM of a physical PE, in particular focusing on the Spanish market (OMEL). The paper is organized as follows: Section 2 reports the results of a statistical modeling performed over real market data (OMEL). The analysis mainly focus on sellers population decision properties. Section 3 describes the Genoa artificial power exchange (GAPEX) and the obtained results in the presence of clearing house mechanism [1]. Section 5 draws conclusions.

Section snippets

Statistical modeling of the market

The OMEL DAM collects all the offers for the 24 h of the following day during the same session and for every hour it determines a clearing price matching the two curves of supply and demand. Generally speaking, stylized facts have been highlighted for the energy spot-price time-series, e.g., mean-reverting behavior, heteroscedastic phenomena, strong seasonal components [2], [3], [4].

In particular, due to the OMEL properties (i.e., a set of 24 offers at once), considering a set of 24 separated

The Genoa artificial power exchange

Recent years have been characterized by an increasing interest on modeling tools for power exchanges, due to the deregulation of production and dealing of electricity in Europe. Different approaches have been adopted by the scientific community: econometric [2], [4], game theory, agent-based [5], [6], etc. The GAPEX has been built with the purpose to emulate a real physical PE with system marginal price (SMP), adopting an heterogenous agent-based system.

In the present version, GAPEX offers the

Conclusion

In this paper, a multi-agent electricity market modeling has been proposed. The Genoa Artificial Power Exchange (GAPEX) is characterized by a fixed number of the participant agents for each negotiation and by a finite total power, that is distributed among sellers. In the computational experiments, the Day Ahead Market session of OMEL (The Spanish Power Exchange) is modeled with a representative buyer based on the real demand and sellers based on statistical modeling. A clearing-house mechanism

Acknowledgements

This work is partially supported by the University of Genoa, the University of Trieste and the Italian Ministry of Education, University and Research (MIUR) under grants FIRB 2001 and COFIN 2004.

References (6)

  • H. Mendelson

    Market behavior in a clearing house

    Econometrica

    (1982)
  • E. Guerci, S. Ivaldi, P. Magioncalda, M. Raberto, S. Cincotti, Modeling electricity spot prices using an arma-garch...
  • A. Léon, A. Rubia, Comportamiento del precio y volatilitad en el pool eléctrico español, Working Paper,...
There are more references available in the full text version of this article.

Cited by (0)

View full text