Abstract
This paper provides a different approach for electricity price forecast from risk management point of view. Making use of neural networks, the methodology presented here has as main concern finding the maximum and the minimum System Marginal Price (SMP) for a specific programming period, with a certain confidence level. To train the neural network, probabilistic information from past years is used. This approach was developed with the objective of integrating a decision-support system that uses Particle Swarm Optimization (PSO) to find the optimal solution. Results from realistic data are presented and discussed in detail. © 2005 ISAP.

Abstract
In a liberalized electricity market, participants have several types of contracts to sell or buy electrical energy. Increasing electricity markets liquidity and, simultaneously, providing to market participants tools for hedging against spot electricity price were the two main reasons for the appearance of those types of contracts. However, due to the payoff nonlinearity characteristic of those contracts, deciding the optimal portfolio that best adjusts to their necessities becomes a hard task. This paper presents an optimization model applied to optimal contract allocation using Particle Swarm Optimization (PSO). This optimization model consists on finding the portfolio that maximizes the electricity producer results and simultaneously allows the practice of the hedge against the volatility of the System Marginal Price (SMP). Risk management is considered through the consideration of a mean-variance optimization function. An example for a programming period is presented using spot, forward and options contracts. PSO performance in such type of problems is evaluated by comparing it with the Genetic Algorithms (GA).

Abstract
It is known that Cournot game theory has been one of the theoretical approaches used more often to model electricity market behavior. Nevertheless, this approach is highly influenced by the residual demand curves of the market agents, which are usually not precisely known. This imperfect information has normally been studied with probability theory, but possibility theory might sometimes be more helpful in modeling not only uncertainty but also imprecision and vagueness. In this paper, two dual approaches are proposed to compute a robust Cournot equilibrium, when the residual demand uncertainty is modeled with possibility distributions. Additionally, it is shown that these two approaches can be combined into a bicriteria programming model, which can be solved with an iterative algorithm. Some interesting results for a real-size electricity system show the robustness of the proposed methodology. © 2005 Cambridge University Press.

Abstract

Abstract
Most optimization and decision problems in power systems include integer or binary variables, leading to combinatorial problems. In this paper, several approaches using metaheuristics and genetic algorithms are presented that deal with real problems of the power industry Most of these methodologies are now implemented in distribution management systems (DMS) used by several utilities.

Abstract
An artificial neural network concept has been developed for transformer fault diagnosis using dissolved gas-in-oil analysis (DGA). A new methodology for mapping the neural network into a rule-based inference system is described. This mapping makes explicit the knowledge implicitly captured by the neural network during the learning stage, by transforming it into a Fuzzy Inference System. Some studies are reported, illustrating the good results obtained.