Transactions evaluation and unit commitment in a bilateral electricity market using genetic algorithms

Short term optimal operation planning and the evaluation of multiple power transactions (scheduled contracts for power exchange at a specific time int erval) in a deregulated environment based on a bilateral market model is a complex optimization problem for the planning personnel of electric utiliti es. With regard to time dependent constraints and costs in unit commitment (UC), the transaction evaluation problem and UC are coupled and therefore m ust he solved simultaneously. In this paper a new solution method based on genetic algorithm (GA) is presented. Due to the flexibility of GA, it is po ssible to select optimal transactions within the UC algorithm. The basic id ea of the new method is to find an optimal load curve by means of an evolut ionary process using an optimal combination of proposed transactions simult aneous with optimal scheduling of generation units. The capabilities of thi s approach are illustrated by means of a 26-unit system with a set of propo sed transactions that includes exchange with pump storage units and various forms of transactions such as peak, off-peak and base load.