Multi-objective intelligent detailed design for prefabricated composite slabs using two-level multi-population co-evolution algorithm

Abstract

The prefabricated composite slab (PCS) is an essential horizontal component of precast buildings. The detailed design process for PCS is extremely complex and challenging due to the need for considering multi-disciplinary and cross-stage collaborations. Traditionally, rule-based methods for PCS design are time-consuming and labor-intensive to provide high-quality and error-free solutions. Therefore, an intelligent detailed design framework is developed to provide necessary manufacturing information for each PCS and its rebar mesh. Specifically, a two-level multi-population co-evolution algorithm (MPCEA) is proposed to solve the high-dimensional optimization problem associated with big-scale PCS design. In the rebar layout, a non-uniform sampling strategy is utilized to generate the high-quality initial population, and a greedy selection method is utilized to obtain the optimal co-evolutionary solutions. The first-level adjusts the positions and dimensions of all PCSs to reduce the number of slab specifications and quantities of slabs, and the second-level ensures collision-free rebar meshes with fewer specifications. Two different examples are illustrated to validate the feasibility of the proposed framework. The experimental results demonstrate that the multi-population differential evolution (MPDE) and multi-population grey wolf optimization (MPGWO) methods perform better compared to other methods.