Comparison of linear, nonlinear, and network flow programming techniques in fuel scheduling

Abstract

The objective of fuel scheduling is to minimize the operating cost over a given time period while satisfying several constraints including fuel supply, fuel inventory, total power requirement, and generating unit power limits. This paper presents a comparison of network flow programming (NFP) the out-of-kilter method, the linear programming (LP) revised simplex method, and the nonlinear programming (NLP) generalized reduced gradient method in order to solve the fuel scheduling problem. For application in NFP and LP, a separable programming technique is used to linearize the objective function and the constraints. Results based on sample data obtained from Omaha Public Power District are presented to demonstrate the differences between NFP, LP, and NLP application to this problem.