Deterministic Parameter Control Applied to the Grouping Genetic Algorithm with Controlled Gene Transmission

An experimental study was performed where de-terministic parameter control is applied to the crossover and mutation ratios of the Grouping Genetic Algorithm with Controlled Gene Transmission (GGA-CGT) considering several control functions with different characteristics in order to solve the offline One-Dimensional Bin Packing Problem (1-D BPP). Such control is mainly based on sinusoidal behavior to allow the algorithm to experience small variations in the parameter to control, and another control scheme based on the same sinusoidal behavior was designed to experiment smaller and larger parameter step size values even in the last stages of the run. An additional form of control is also included based on linear growth. The obtained performance is compared against the tuned version of the algorithm. Obtained results show that the GGA-CGT could benefit from the sinusoidal control scheme under small variations for the crossover ratio and from the linear control scheme for the mutation ratio. Lastly, the control schemes that showed the best results were tested in the Grouping Genetic Algorithm with Intelligent Heuristic Strategies (GGA-IHS) designed to solve the Parallel-machine scheduling problem with unrelated machines and makespan minimization
$(R\Vert C_{\max})$
showing promising results.