Combinatorial optimization approach for the efficient reuse of RC components

Abstract

The reuse of reinforced concrete components from deconstructed buildings offers a promising approach to reduce the environmental impact of new constructions. However, it represents a complex combinatorial optimization problem to efficiently place the available modules, which vary in geometry and load-bearing capacity, into new structures while maximizing their utilization. This paper proposes a two-stage optimization method to enable the reuse of arbitrary reinforced concrete modules. First, an agent-based model is employed to rapidly explore feasible geometric combinations of modules and preselect suitable placements based on a target span length. Second, metaheuristic optimization algorithms, namely Simulated Annealing and Tabu Search, are adapted to maximize the utilization of the modules' load-bearing capacity while ensuring global structural integrity. The methods are demonstrated on a case study of assembling a three-span continuous beam from a sampled construction kit of 100 reinforced concrete modules with varying cross-sectional properties and material parameters. The results show the agent-based preselection effectively finds viable geometric combinations, while the metaheuristics converge on optimised module placements with up to 88% utilization on average. The proposed approach provides a computational framework to enable the direct reuse of structural concrete components, supporting the design of low-carbon circular buildings.