Multiple unmanned aerial vehicle (multi-UAV) cooperative air combat, which is an important form of future air combat, has high requirements for the autonomy and cooperation of unmanned aerial vehicles. Therefore, it is of great significance to study the decision-making method of multi-UAV cooperative air combat since the conventional methods are challenging to solve the high complexity and highly dynamic cooperative air combat problems. This paper proposes a multi-agent double-soft actor-critic (MADSAC) algorithm for solving the cooperative decision-making problem of multi-UAV. The MADSAC achieves multi-UAV cooperative air combat by treating the problem as a fully cooperative game using a decentralized partially observable Markov decision process and a centrally trained distributed execution framework. The use of maximum entropy theory in the update process makes the method more exploratory. Meanwhile, MADSAC uses double-centralized critics, target networks, and delayed policy updates to solve the overestimation and error accumulation problems effectively. In addition, the double-centralized critics based on the attention mechanism improve the scalability and learning efficiency of MADSAC. Finally, multi-UAV cooperative air combat experiments validate the effectiveness of MADSAC.

(1) Background: Achieving harmonious human–land relations is one of the key objectives of sustainable urban–rural development, and the degree of decoupling of permanent population levels from changes in construction land use is an important factor in related analyses. Due to the existence of huge urban–rural differences, rethinking China’s human–land relations from the perspective of integrating urban and rural areas is of great value for the advancement of high-quality urban–rural development. (2) Methods: By studying the lower reaches of the Yangtze and Yellow Rivers of China, and based on data from the second and third national land surveys of China, this paper analyzes the spatio-temporal evolution of urban and rural population, construction land use, and human–land relations from 2009 to 2019 using exploratory spatial data analysis (ESDA) and a decoupling model; in addition, this paper proposes a differentiated zoning management strategy and establishes a new framework that integrates evolutionary patterns, human–land relations, spatial effects, and policy design. (3) Results: The geographic distribution patterns of urban and rural population and construction land use remained stable over time, with high levels of spatial heterogeneity, agglomeration, and correlation. Changes in urban and rural population levels and construction land use are becoming increasingly diversified and complex, with both increases and reductions existing side by side. Based on a Boston Consulting Group matrix, the evolution patterns of urban and rural population and construction land use are divided into four types, referred to as star-cities, cow-cities, question-cities, and dog-cities. Over the time period examined in this paper, the spatial autocorrelation of urban land evolution patterns turned from negative to positive; however, that of rural land, as well as those of urban and rural population evolution patterns, were statistically insignificant. Urban human–land relations are coordinated, in general, and are mostly in a state of either weak decoupling or expansive coupling. In contrast, rural human–land relations are seriously imbalanced, and most of them are in a state of strong negative decoupling. Human–land relations are dominated by regressive changes in urban areas but remain unchanged in rural areas. Cold- and hot-spot cities are concentrated in clusters or in bands, forming a core-periphery structure. The formation and evolution of the decoupling relationship between construction land use and permanent population are the results of multiple factors, including urbanization, industrialization, globalization, and government demand and policy intervention. The interaction effects between different factors show bifactor enhancement and nonlinear enhancement, with complex driving mechanisms and large urban–rural differences. It should be highlighted that the influence intensity, operation mechanism, and changes in the trends for different factors vary greatly. Urbanization rate, gross domestic product, and government revenue are key factors that exert a strong direct driving force; international trade, foreign direct investment, and per capita GDP are important factors, while the remaining factors are auxiliary factors that remain heavily dependent on interaction effects. (4) Conclusions: To further transform human–land relations from imbalanced to coordinated, we divide the study area into four area types based on the concept of urban–rural community: urban and rural intensive policy areas, urban intensive policy areas, rural intensive policy areas, and urban and rural controlled policy areas. Furthermore, we put forward suggestions on the differentiated management of land use for the four types of policy areas.

No abstract

No abstract