Exploring close-optimal solutions for the Time Constrained Scheduling Prolem in project management

This paper presents a new approach for the Time Constraint Project Scheduling Problem (TCPSP). A two stage heuristic algorithm was developed for this problem. During the first stage a genetic algorithm using as fitness function, a function composition of max resource usage and the differences between actual and desirable resource usage, is executed to get a set of solutions. The second stage consists of choosing the best chromosomes and moving the solution´s set start times ±1 to randomly generate another set of solutions. Experimental results of the two stage approach are presented and compared to the single genetic algorithm results.