Dynamic Programming Approach to Unit Commitment

Abstract

A field-proven dynamic programming formulation of the unit commitment problem is presented. This approach features the classification of generating units into related groups so as to minimize the number of unit combinations which must be tested without precluding the optimal path. Programming techniques are described which maximize efficiency. Considerations are discussed which determine when generating units must be evaluated and when they may be ignored. The heuristic procedures described in this paper are concerned with supplying all apriori information to the program thereby minimizing its execution time. Results are presented from field testing on a medium size utility. Composite generating unit formulation is described for the economic allocation of constrained fuel to a group of units.