A partition approach algorithm for nonconvex economic dispatch

This paper develops a partition approach algorithm (PAA) to solve economic dispatch with non-smooth fuel cost. The PAA developed integrates the multi-section divided and interval elimination method to formulate the searching algorithm. The populations produced of PAA are model by parameter interval and divided sections to calculate the combinations of effective factors and their selected rules. The analysis of results uses a signal-to-noise ratio (S/N ratio) to aid in the determination of the offsprings. Effectiveness of the PAA is demonstrated on three example systems with a nonconvex function and compared to that of SQP, SA, GA, and EP. The results show that the proposed method are more effective than other previously developed evolutionary computation algorithms and have a great potential for other applications in power systems.