Multiobjective Ant Colony System Algorithm for Component-Level Construction Schedule Optimization

Abstract

Automatic generation of construction schedules has emerged as a key solution to address the inefficiencies and instabilities arising from the over-reliance on empirical judgment. However, traditional construction scheduling has been predominantly limited to regional levels, inadequately addressing the lean construction requirements of component-based prefabricated steel frame (PSF) structures. To bridge this gap, this study formulates an optimization model for the component-level resource-constrained project scheduling problem for PSF structures (C-RCPSP-PSF), which realizes the automatic extraction of precedence relationships from building information modeling three-dimensional (BIM 3D) models and the minimization of construction duration, costs, and carbon emissions. To address the C-RCPSP-PSF model, a novel multiobjective ant colony system (MOACS) algorithm is developed that utilizes three distinct colonies to individually tackle the objectives and combines taboo lists and global archives to enhance the search. Experimental results show the superior convergence and diversity of the MOACS over that of other competitive multiobjective optimization algorithms in solving the proposed model.