Abstract: When spot market price and customer demand are correlated, the paper investigates a two-stage procurement risk management model based on portfolio contracts and spot market with limited capacity so as to maximize the retailer's expected profit.Firstly we use back-ward induction to derive the retailer's optimal strategy in the second stage, then employ the standard perturb argument to provide with the optimality property of consecutive active contracts.Secondly, we transfer the original model into a shortestmonot one path problem and propose a dynamic programming algorithm to obtain the retailer's optimal procurement strategy.Finally, we employ a numerical example to study the im pacts of correlation and capacity in the spot market on the optimal procurement strategy, and the conclusion are drawn: (1) given the capacity of the spot market, the optimal reservation amounts of active contracts raise as the correlation increases, moreover the intensity of increase depends on the contract's flexibility; (2) given the correlation, as the capacity of spot market increases, the retailer should decrease the optimal reservation amounts of active contracts as well as active contract with the lowest execution cost.

Key words: procurement risk management, option contract, spot market, shortest-monotone path, dynamic programming