Backward Dynamic Programming Approach to Solve the Unit Commitment Problem Based on a Day Load Curve

Abstract

Unit Commitment is generally used to schedule the units so as to minimize the cost of the production for the total system across the scheduled time horizon under generator operational constraints. The Unit Commitment problem is nonlinear, and a mixed integer type combinatorial optimization problem. We can obtain the global optimal solution by complete listing which cannot be possible for large power systems due to its requirements for excessive computational time. Unit Commitment presents a methodology to reduce cost and improve the system reliability. Unit Commitment is a dynamic process and power production planning changes as per loads and topologies of the network. The problem is solved by Backward Dynamic Programming approach which solves the dynamic problem in 12 stages. This paper deals with a Unit Commitment problem that is solved using Backward Dynamic programming without time constraints and outcomes show minimum cumulative total cost for operating 4 units in 12 stages for a 24-hour horizon based on a load curve of a day