The advent of competitive markets confronts each producer with the problem of optimally allocating his energy/capacity so as to maximize his pro$ts. The multiplicity of auctions in electricity markets andthe nontrivial constraints imposedby technical andbid ding rules make the problem of crucial importance andd i/cult to model and solve. Further di/culties are represented by the dynamic and stochastic natures that characterize the decision process. We formulate the problem as a multi-stage mixed-integer stochastic optimization model under the assumption that the seller is a price taker. We validate the e2ectiveness of the proposed model on a representative test problem.