Skip to Main Content
The applicability of genetic algorithms (GA) to the generator maintenance scheduling (GMS) problem with modified genetic operators (MGO), such as string reversal and reciprocal exchange mutation (REM) is demonstrated. The main contribution is the use of 'probabilistic production simulation' (PPS) with an equivalent energy function method, which outperforms other methods in terms of computation time and accuracy. The performance of the algorithm has been tested on 5- and 21-unit test systems with integer encoding, binary for integer encoding, and real encoding. The GMS problem is solved to minimise the expected energy production cost (EEPC) and maximising the reserve objectives under a series of constraints. Results are compared with solution by conventional methods. This paper places in proper perspective the effect of MGO, with an explicit case study and simulation results. It is placed in evidence that only integer coding GA finds the global optimum solution, irrespective of the nature of the objective function and system size. Faster convergence is enhanced with the implementation of MGO for integer GA only.