A systematic review identifying sustainability and uncertainty gaps in construction optimization

This systematic review synthesizes 123 linear programming and mixed-integer programming (LP/MILP) optimization studies in construction management (1985–2024), and assesses the extent of sustainability integration and uncertainty treatment across seven application domains. Two principal findings emerge. First, a substantial sustainability gap persists: while all reviewed models include economic objectives or constraints, 89% of studies optimize exclusively for economic outcomes, only 11% incorporate environmental metrics alongside economic objectives, and social metrics are entirely absent (0%) despite 37% of the literature employing multi-objective methods capable of accommodating additional dimensions. Second, an uncertainty paradox characterizes this strand of practice: although 53% of models remain purely deterministic, the domains most exposed to operational variability, Site Layout and Logistics (91% deterministic) and Resource Allocation (78% deterministic), show the weakest adoption of stochastic or robust methods. These patterns reflect problem framing shaped by data availability, disciplinary boundaries, and contractual structures rather than technical limitations. The review outlines a research agenda centered on embedding social metrics, extending uncertainty-aware formulations, and pursuing cross-domain integration, and argues that realizing this potential will also require institutional conditions including contract structures, procurement criteria, and interdisciplinary incentives that reward sustainability-oriented optimization. Because the search strategy specifically targets LP/MILP formulations, these findings characterize the mathematical-programming strand of construction optimization and should not be generalized to all computational approaches.