A Techno-Economic Decision Support Framework for Construction Swarm Robotics: An Analysis of Value-Blind vs. Value-Aware Control Algorithms

While the adoption of swarm robotics is accelerating in the dynamic and uncertain environment of the construction industry, conventional path-planning algorithms based on shortest physical distance induce severe bottlenecks and schedule delays during large-scale swarm operations. To expand the control paradigm of smart construction automation by integrating ‘physical shortest-distance optimization’ with ‘financial value maximization,’ this study proposes a ‘Cost-Topology Map’ that translates site uncertainties into economic risk costs. An evaluation pipeline is established by integrating Techno-Economic Assessment (TEA) techniques with a Quantitative Technology-Organization-Environment (TOE) Framework, directly linking engineering simulation metrics (makespan, success rate) to corporate financial cash flows (net profit, ROI). A total of 875,000 macroscopic, scenario-based Monte Carlo simulations were conducted to evaluate the techno-economic consequences of the literature-informed behavioral parameterizations representing seven major swarm control paradigms (Random Walk, APF, Voronoi, Flocking, Digital Pheromone, Market-based, and MARL) under varying sensor noise levels (5–30%) and swarm sizes (50–1000 units). The analysis reveals that only the value-aware Market-based and MARL models preemptively secured high-value tasks, achieving meaningful profitability up to the 500-unit scale (N = 500) and indicating a comparative techno-economic advantage over value-blind classic profiles within the modeled framework. Conversely, in a 1000-unit scale-up environment (N = 1000), exponentially increasing fleet operating costs and bottleneck delay penalties drove even the value-aware MARL profile beyond its break-even threshold into a financial deficit, quantitatively demonstrating the financial impact of the ‘Scale-up Trap’. Furthermore, under complex extreme conditions with severe sensor noise (30%), the Market-based profile experienced a pronounced financial deficit (−93.4 million KRW) due to distorted bidding information, whereas the MARL profile demonstrated stronger downside-risk mitigation by limiting the deficit (−64.5 million KRW) under the parameterized nonlinear policy assumptions. This study holds academic significance by proposing an objective evaluation model that supports practical executive decision-making when adopting large-scale swarm robotics, translating engineering performance gaps into projected financial outcomes.