Multi-Objective Optimisation of Arable Land Allocation Within Environmental Carrying Capacity

Widespread abandonment of arable land, extensive cultivation practices, accumulation of fertiliser and pesticide-derived pollution jointly threaten sustainable use of regional arable land resources. This study addresses the allocation of a fixed regional arable land endowment between two operating entities with distinct behaviour: land contractors, who lease land, pursue profit, apply higher agrochemical inputs and may leave part of the leased area idle, and farmers, who cultivate their own land with higher yields and lower emissions. A tri-objective programming model is constructed that maximises the utilisation rate of arable land, maximises total grain yield and minimises the environmental pollution load, subject to the regional environmental carrying capacity, a government subsidy budget and a tiered penalty on idle land. Fertiliser and pesticide non-point-source pollution and crop-straw disposal: open burning versus return to the field, with the associated air pollutants, water pollutants and nitrogen losses, are converted into standard pollution equivalents following the NY/T 4157-2022 guideline. The model is solved with the NSGA-II algorithm, which outperforms GA and MOPSO on generational distance (0.0225), hypervolume (0.8832) and spread (0.0201). Sensitivity analysis shows that the subsidy coefficient for contractors is the dominant lever: it raises land utilisation and grain yield most strongly, but beyond a threshold it drives the largest increase in pollution load, while the waste-emission coefficients of the two entities jointly amplify environmental pressure. Based on the Pareto solution set, three differentiated regulation schemes: low-pollution priority, high-yield priority and balanced development, are proposed for coordinating subsidies, idle-land penalties and agrochemical application limits.