Scheduling multiple job problems with guided evolutionary simulated annealing approach

This paper reports on an investigation of whether a special type of evolutionary programming named guided evolutionary simulated annealing (GESA) might be used effectively for dealing with scheduling tasks. The GESA approach allows many candidate solutions to be `alive´ at the same time. There is local competition and global competition and more and more search resources are guided into promising regions. Simulated annealing avoids entrapment in local minima. Two examples of multiple job scheduling were investigated. Results obtained with GESA were superior to those obtained with a simulated annealing approach described in prior literatures