Three-phase CPLEX with quota auction mechanism for dedicated resource-constrained multi-project scheduling problem with calendar constraints

Purpose In the realistic multi-project scheduling, resources are not always shared among multiple projects, nor are they available to perform activities throughout the planning horizon. Besides, according to construction technology, some architectural jobs cannot be interrupted for any reason. However, these characteristics of resources and activities have not been fully studied, which may lead to the reduction of engineering quality and the failure of scheduling work. Therefore, this paper aims to model a multi-project scheduling problem with the above characteristics and provide an effective method to meet the actual needs of the construction industry. Design/methodology/approach A three-phase CPLEX with quota auction mechanism (TPCP–QAM) is developed to solve this problem, which significantly improves the solving performance of CPLEX by adjusting the search strategy and implementing a distributed procedure. In this approach, resources are dedicated to individual projects through a global coordination mechanism, while each project is independently scheduled by a local scheduling algorithm. Findings (1) For the proposed problem, CPLEX 2019's default search strategy. (Auto) is far inferior to another search strategy (Multi-point) in optimizing the project total cost and average resource capacity. (2) Compared with other two algorithms, TPCP–QAM has obvious advantages in the multi-project total cost (MPTC) and CPU time, especially for large-size instances. (3) Even though the number of non-working days may not be changed for the protection of labor resources, managers can reduce MPTC or shorten the multi-project total makespan (TMS) by appropriately adjusting the distribution of non-working days. Originality/value This paper fulfils an identified need to investigate how to complete a multi-project portfolio with the minimum cost while ensuring engineering quality under a practical multi-project scheduling environment.