A multi-objective optimization approach to project scheduling with resiliency criteria under uncertain activity duration

• Published in: Annals of operations research 2020-02, Vol.285 (1-2), p.161-196
• Main Authors: Torabi Yeganeh, Farnaz, Zegordi, Seyed Hessameddin
• Format: Article
• Language: English
• Subjects: Buffers; Business and Management; Case studies; Classification; Combinatorics; Combined cycle power generation; Computer simulation; Critical path; Disruption; Electric power generation; Floating structures; Genetic algorithms; Industrial project management; Mathematical models; Mathematical optimization; Methods; Multiple objective analysis; Operations research; Operations Research/Decision Theory; Optimization; Project management; Reliability; Resilience; Resource allocation; Resource scheduling; S.I. : Project Management and Scheduling 2018; Schedules; Scheduling (Management); Sorting algorithms; Theory of Computation; Uncertainty
• ISSN: 0254-5330; 1572-9338
• DOI: 10.1007/s10479-019-03375-z

Uncertainty is one of the main parts of the project management environment that can strongly affect the project objectives and cause unpredictable delays. This study presents a multi-objective optimization approach for constructing resilient project schedules under resource constraints to cope with uncertain activity durations. In this paper, the concept of resilient project scheduling is defined to measure the ability of schedules to deal with duration disruption. Since the direct evaluation of resiliency is computationally complicated and time-consuming, a new surrogate resilience measure is introduced. The proposed resiliency criteria measure the floating of activities and the risks associated with the completion of the project. Furthermore, a new model based on a combination of time buffer and float allocation approach is developed. To extend existing project scheduling models with uncertainty, general precedence relationships between activities have been considered. To validate the proposed approach, the construction project of a combined cycle power plant is used as a case study. Due to a large number of project activities in this case study, the non-dominated sorting genetic algorithm (NSGA II) has been used to solve the problem. The results of solving the mathematical model using the proposed method are assessed through extensive simulation experiments and compared with those of the baseline schedule. The results show that by taking the proposed resiliency measure and the optimal allocation of buffer time to activities, the project completed at the same duration with higher reliability.