A Cuckoo Search Algorithm for Scheduling Multiskilled Workforce

Based on the situation of the construction and software industry, this paper presents a multi-skilled human resource scheduling model which considers the skill and the salary of the workforce at different levels. It introduces a novel cuckoo search (CS)-based approach to optimize the objective function which combines the minimization of the makespan and the project cost. The proposed approach takes the priority coding scheme to code for each job, and adopts the serial schedule generation scheme to sequentially allocate the start time and resources for one job with highest priority at a time. To match the discrete nature of the solution, the Lévy flight operator of CS algorithm, which is originally suitable for continuous variable, has been modified. The CS-based approach has been implemented by using MATLAB R2010b and tested by a project taken from the literature, which contains 18 jobs and 6 multi-skilled workers. The results yielded by the CS-based approach have been found to significantly outperform genetic algorithm. This demonstrates that the cuckoo search algorithm based approach is suitable for scheduling the complicated resource constrained project with multi-skilled workforces.