To address the challenge of multi-objective collaborative optimization in low-carbon landscape design, this study constructs an optimization framework that integrates landscape pattern indices with a multi-objective genetic algorithm. First, key landscape pattern indices—including patch density (PD), edge density (ED), Shannon evenness index (SHEI), mean nearest neighbor distance (MNN), and connectivity (CONT)—are selected to establish an evaluation system for low-carbon landscape spatial structure. Building upon this foundation, the second-generation non-dominated sorting genetic algorithm (NSGA-II) is introduced and enhanced through the incorporation of a differential local search strategy, thereby improving its local optimization capability. Using a typical cold-region urban ecological park in Harbin, Heilongjiang Province, as a case study, energy consumption and cost simulations are performed with Design Builder software. Experimental results demonstrate that the proposed algorithm can effectively balance multiple conflicting objectives, such as carbon emissions, construction costs, and resident satisfaction, yielding a well-distributed Pareto-optimal solution set. In tests on the ZDT benchmark functions, the spacing (SP) metric achieves an optimal value of 0.07. The optimized design reduces the park’s construction cost to 208,000 yuan, lowers annual CO₂ emissions to 2180 tons, and significantly improves all resident satisfaction indicators. This study provides a systematic and effective methodology for the quantitative assessment and multi-objective collaborative optimization of low-carbon landscapes.
Ying Pan (2026) studied this question.