A robust multi-project scheduling problem under a resource dedication-transfer policy

doi:10.1007/s10479-024-05854-4

2024-02-21

PRE

DOI：10.1007/s10479-024-05854-4

原文链接

原文求助

摘要

En 中文

In multi-project management, effectively allocating limited resources to projects is crucial for ensuring the success of the project portfolio and timely delivery. There are two approaches to managing renewable resources based on project execution environment and resource characteristics: resource sharing and resource dedication. This study explores a novel resource dedication-transfer policy, where renewable resources are dedicated to each project during execution and can be transferred to other projects once the corresponding project is completed. Additionally, real-world multi-project scheduling often faces uncertainties, with the most common issue being varying durations of activities. To address this, a hierarchical multi-objective optimization model is proposed under the resource dedication-transfer policy. This model aims to allocate dedicated resources to the project portfolio at a tactical level and schedule individual activities at an operational level. The specific objectives include minimizing the total weighted tardiness of all projects, minimizing the total fluctuating costs of resources, and maximizing solution robustness in the presence of activity duration variabilities. The proposed solution methodologies include an adjusted large neighborhood search (ALNS) and a customized NSGA-II algorithm. Both algorithms employ a hybrid coding scheme of project-buffer-resource-activity list to represent feasible solutions. Specially, the ALNS introduces new destroy-repair neighborhood operators and a hypervolume-based selection criterion to enhance its performance. Experimental results demonstrate that while both algorithms have their advantages for small-scale instances, the ALNS is more effective for large-scale instances. Finally, the derived Pareto solutions from the proposed method are further evaluated through a simulation of multi-project execution under different levels of activity duration variabilities.

Keyword:

Multi-project scheduling

Resource dedication

Resource transfer

Duration uncertainties

Multi-objective optimization

Adjusted large neighborhood search