Evaluating zoning-based lift stopping strategies under stochastic worker arrivals in super-tall construction using agent-based simulation

Managing vertical transportation during peak periods is a critical challenge in super-tall construction projects, directly affecting labor productivity, schedule reliability, and site operations. Existing studies have primarily relied on deterministic optimization to improve lift performance, but they offer limited insight into how alternative stopping strategies perform under realistic, time-varying worker arrivals. This study develops an agent-based simulation framework that explicitly models lift cars and workers as interacting agents and evaluates zoning-based lift operations under stochastic crew-arrival scenarios. Using a 123-storey ‘J’ skyscraper as a case study, the model is first validated against a deterministic optimization baseline and then extended to represent probabilistic worker arrivals fitted from empirical data. Within this stochastic framework, three stopping strategies—Top-Floors-First, Sequential Bottom-Up, and High-Demand-Floors-First—are systematically compared. The results show that the proposed Top-Floors-First strategy consistently outperforms the other two strategies under stochastic-arrival conditions, reducing total operating time by 12–20% and average worker waiting time by 6–19%. These improvements arise from prioritizing upper-floor destinations, thereby reducing backtracking and shortening lift cycles. By capturing probabilistic arrival patterns and micro-level lift–worker interactions, the proposed approach advances construction logistics modeling and provides a practical decision-support tool for planning robust lift operations under uncertainty.