Modelling the effect of multi-stakeholder interactions on construction site layout planning using agent-based decentralized optimization

• Published in: Automation in construction Vol. 107; p. 102927
• Main Authors: Song, Xiaoling, Peña-Mora, Feniosky, Shen, Charles, Zhang, Zhe, Xu, Jiuping
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01-11-2019; Elsevier BV
• Subjects: Agent-based modelling; Construction materials; Construction sites; Decentralized optimization; Decisions; Environmental impact; Layouts; Logistics; Mathematical models; Multi-stakeholder interaction; Particle swarm optimization; Planning; Safety management; Security planning; Site layout; Stakeholders; Supply channel design; Site layout; Security planning; Agent-based modelling; Supply channel design; Decentralized optimization; Multi-stakeholder interaction
• ISSN: 0926-5805; 1872-7891
• DOI: 10.1016/j.autcon.2019.102927

Construction site layout planning (CSLP) is important to ensure that cost, time, quality and environmental impacts are strictly controlled. Even though extensive research has been conducted on CSLP problems, the effect of the multi-stakeholder interactions on optimal site layout plans has been less studied. Therefore, this paper proposes an agent-based decentralized bi-level mathematical model that incorporates decisions from three stakeholders, to examine the CSLP multi-stakeholder interactions. To solve the model, a particle swarm optimization (PSO)-based solution approach is developed to generate the CSLP, construction material logistics planning (CMLP) and security planning (SP) solutions in an integrated CSLP-CMLP/SP system. Specifically, the stakeholders are modelled as individual agents with respective decisions and objective functions, and the inter-dependencies between the layout planner and the other two stakeholders (i.e. the logistics planner and the safety manager) are simultaneously incorporated and dealt with using the bi-level model. The multi-stakeholder interactions are realized by iteratively updating the individual decisions to determine an optimal integrated solution in which the potential stakeholder conflicts in the construction stage are minimized. A case study is given to validate the practicality of the designed model in addressing the CSLP-CMLP and CSLP-SP interactions. The contributions of this research are its applicability to model the effect of multi-stakeholder interactions in agent-based decentralized models, the design of a PSO-based solution approach to achieve an integrated CSLP-CMLP/SP solution, and the provision of academic and practical guidance for decision makers to further enrich global decentralized CSLP research.
•An agent-based decentralized model is proposed to study the effect of multi-stakeholder interactions on CSLP.•The system stakeholders are modeled as individual agents with their respective decisions and objective functions.•Agent-based modelling and bi-level programming are integrated to model the mutual interactions.•The paper enriches CSLP research by guiding academic and practical decision makers in a decentralized and coordinated manner.