Generative layout optimization of BIM schemes for urban renewal districts under CIM environmental constraints based on SA and GA algorithms

During the advancement of urban renewal projects, Building Information Modeling (BIM) layout plans for renewal areas must balance optimal and reasonable comprehensive benefits with planning efficiency. Over-reliance on manual, experience-based planning not only limits the scope of planning for renewal areas but also undermines the effectiveness of urban renewal decisions. This paper proposes a generative layout optimization method for BIM schemes in urban renewal areas under City Information Modeling (CIM) constraints. Considering economic agglomeration power and uniformity index, the comprehensive benefits are analyzed, and the problem is mathematically formulated to maximize economic returns from building combinations and site arrangements. Using CIM-based environmental information, soft and hard constraints are established. Based on the characteristics of the optimization objective, this study innovatively designs an adaptive genetic simulated annealing algorithm. The algorithm employs an adaptive linear scaling adjustment strategy for fitness on top of the traditional genetic algorithm, which prevents premature convergence while suppressing the homogenization of generative layout optimization outputs. Simultaneously, the probabilistic jump-like nature of the simulated annealing algorithm is leveraged to avoid the risk of local optima. In each generation of the genetic algorithm, neighborhood perturbations are applied to the population individuals to further enhance the algorithm’s performance. The Metropolis criterion determines whether to accept new individuals, thereby balancing global exploration and local exploitation. This approach addresses the shortcomings of the single genetic algorithm - namely, its tendency to converge prematurely and become trapped in local optima - and achieves global optimization of the layout. Experimental results show that when the maximum iteration condition is met, the proposed method achieves a runtime of 2.51 s, a solution accuracy of 97.56%, and a constraint satisfaction rate of 100% in generative layout optimization. This indicates that the design method responds quickly when applied to the optimization of urban renewal areas, and its results more closely approximate the theoretically optimal layout, which is more conducive to debugging generative layout optimization in planning decision-making. The BIM layout generated for the urban renewal area under CIM environmental constraints after optimization yields a comprehensive benefit of 0.927, an economic benefit of 62,238 yuan, and a rationality level of 86.79%. This indicates that the design method can efficiently approximate the Pareto front of the target optimization in generative layout optimization for urban renewal districts. The optimization results demonstrate synergy in constraint satisfaction, compliance, and economic efficiency, and exhibit strong global search capabilities. The above results validate the effectiveness of the proposed method in preventing premature solutions, suppressing homogenization, and avoiding local optima. Compared with traditional algorithms, it possesses stronger optimization capability and efficiency, and can enhance comprehensive benefits within the renewal area while ensuring the overall rationality and economic viability of the building layout.