In most of the multi-objective multi-skill resource-constrained project scheduling (MO-MSRCPSP) research works, it has been assumed that a staff member is able to exhibit different skills with the same proficiency or expertise. However, this is not true in real life. Usually, a person possessing various skills may be expert in one (or two) skill(s) but may only be moderately trained for performing other skills. The assignments of persons with less-skilled levels have to be kept as low as possible to achieve satisfactory quality targets. Under this motivation, this paper develops a mathematical model for a MSRCPSP with two objectives. In addition to the regular objective of minimizing the makespan, the second objective aims at minimizing the total time elapsed with the less-skilled resource assignments defined as ‘skill divergence span’. A weighted sum teaching-learning-based optimization (TLBO) algorithm is employed to solve this complex problem. In addition to the TLBO, a multi-objective genetic algorithm (GA) is also developed as an alternative metaheuristic for the comparison purposes. The computational results are performed on 36 test instances with varying level of skill factors, resource strength and network complexity. The average % deviation from critical path based lower bound obtained is comparatively lower for the MO-TLBO as compared to the MO- GA. It is 62.17% for the proposed MO-TLBO while for MO-GA its value is 75.11% which shows that TLBO is an effective metaheuristic for solving such problems.

Keywords

Teaching-learning-based algorithm (TLBO), multi-objective multi-skill resource-constrained project scheduling (MO-MSRCPSP)