On the Application of Evolutionary Algorithms with Multiple Crossovers for Solving the Task Scheduling Problem

J. Józefczyk and R. Czaja (Poland)


scheduling, evolutionary algorithms, simulation


The complex problem of task scheduling is considered. It is a generalization of the conventional scheduling problem, which consists in taking into account the movement of executors performing tasks located at stationary workstations. The generalization leads to a new NPhard discrete optimization problem, which is solved using the advanced evolutionary algorithms. Their three versions are considered for different multiple crossover operators. The following multiple crossover operators are applied: single crossover on multiple parents, multiple crossovers on multiple parents and multiple crossovers per couple. The solution algorithms with multiple crossovers and with simple crossover are compared during computer simulation. The quality of the scheduling and the time of computation are the basis for the comparison. The simulation shows the moderate advantage of the solution algorithms with multiple crossovers in comparison with the version with a traditional crossover operator.

Important Links:

Go Back