A multiobjective optimization approach to thermal generating units maintenance scheduling

This paper presents a new approach to preventive maintenance scheduling of thermal generating units in electric power systems. The problem has been formulated as a combinatorial optimization task, with explicit and simultaneous treatment of multiple objectives: minimization of fuel costs, maximization of reliability, and minimization of constraints violations. The optimization is performed by the multiobjective branch and bound algorithm with successive approximations, embedded in an outer loop where in each iteration a new initial solution is generated in a pseudorandom manner. This procedure was the basis for development of the multiobjective maintenance scheduling software package. A realistic example of annual maintenance scheduling of 21 thermal generating units demonstrates the effectiveness of the proposed methodology.