Power generation companies have to meet consumer demand and reliability
criteria. The maintenance of generators needs to be performed on a
regular basis to ensure a reliable supply of electricity. An increase in
repair frequency will result in greater maintenance cost together with
an expected decrease in capital loss due to generator failure. An
optimal maintenance schedule is therefore needed to find the best suited
trade-off and lower the overall operation cost.
Maintenance scheduling is a combinatorial optimization problem for which
an exhaustive search is typically infeasible due to the large size of
the solution space. Metaheuristic optimization techniques are therefore
used to approximate global optimum schedules in finite time. This paper
compares the use of several metaheuristic methods in the solution of the
generator maintenance scheduling problem. In particular, our methods
improve on a previously implemented solution of generator scheduling for
the Power Generation Company of Trinidad and Tobago.
Keywords– generator, maintenance scheduling, tabu
search, simulated annealing, genetic algorithm, memetic algorithm