This thesis presents a general framework for the two-period peak load pricing with a central and a decentralized storage. Decreasing return to scale, multiple generation technology, and partitioning electric load into thermal and electric components are assumed. To elaborate the framework, the network describing the problem is constructed. The equilibrium is obtained by the network equilibrium algorithm. The welfare change resulted from the introduction of storage is a decisive analysis tool. This study also investigates several topics in the peak load pricing problem. Major findings are : First, if the storage cost is reasonable, the introduction of storage makes contributions to the social welfare. Second, a central storage is more effective than a decentralized one. Third, thermal storages currently available to the consumer can not be successfully commercialized owing to their high costs. Fourth, the peak (price) reversal does not occur in the general model.