Modified ant colony algorithm for job shop scheduling problem

In this work, we proposed a modified ant colony algorithm (ACA) for job shop scheduling problem (JSSP) with make-span, and constraints such as machine selection, time lags, and holding times, process, and sequence are taken into account. The two-stage setup of the pheromone update mechanism allows for a combination of local and global pheromone updates. In the first stage, the pheromone is updated locally for each completed process, and after the set iteration conditions have been met, the second stage is entered. To overcome the initial reliance on pheromones in the ACA, the pheromones are initialised using a genetic algorithm (GA). The optimal convergence ratio is obtained through the design of a genetic operator based on the procedure principle to accelerate the convergence effect of the whole algorithm and improve the global searching ability of ACA. Taking an engine company as an example, several simulation experiments are carried out for GA, ACA, and modified ant colony algorithm (MACA) based on the standard dataset to verify the effectiveness of proposed algorithms.