Published October 5, 2021 | Version Author version
Journal article Open

Agent-based retail competition and portfolio optimization in liberalized electricity markets: A study involving real-world consumers

  • 1. LNEG

Description

The liberalization of energy markets brought full competition to the electric power industry. In the wholesale sector, producers and retailers submit bids to day-ahead markets, where prices are uncertain, or alternatively,they sign bilateral contracts to hedge against pool price volatility. In the retail sector, retailers compete to sign bilateral contracts with end-use customers. Typically, such contracts are subject to a high-risk premium—that is, retailers request a high premium to consumers to cover their potential risk of trading energy in wholesale markets. Accordingly, consumers pay a price for energy typically higher than the wholesale market price. This article addresses the optimization of the portfolios of retailers, which are composed of end-use customers. To this end, it makes use of a risk-return optimization model based on the Markowitz theory. The article presents a simulation-based study conducted with the help of the MATREM system, involving 6 retailer agents, with different risk preferences, and 312 real-world consumers. The retailers select a pricing strategy and compute a tariff to offer to target consumers, optimize their portfolio of consumers using data from the Iberian market, sign bilateral contracts with consumers, and compute their target return during
contract duration. The results support the conclusion that retail markets are more favourable to risk-seeking retailers, since substantial variations in return lead to small variations in risk. However, for a given target return, risk-averse retailers consider lower risk portfolios, meaning that they may obtain higher returns in both favourable and unfavourable situations.
 


 

Files

Manuscript_watermark.pdf

Files (946.5 kB)

Name Size Download all
md5:dd8cda5c2f121b36c2deba6525082237
946.5 kB Preview Download