| With the construction of ecological civilization and the promotion of new industrialization and urbanization,China’s economic and social development has entered a new stage of high-quality development.Exploring and realizing the optimal allocation of land and space resources is the core issue of establishing a scientific and orderly spatial planning system.Land resources are the core resources of land and space,and also the basis for human beings to carry out all economic and social activities.However,with the rapid advancement of economic and social globalization and human development urbanization,the contradiction between supply and demand of land resources is becoming increasingly serious.How to use scientific theories and methods to guide land spatial planning and promote the sustainable use of land resources has become a hot issue in many related fields.The optimal allocation of land use is one of the core contents of land spatial planning.Around the solution of the optimal allocation of land use,relevant scholars at home and abroad have carried out extensive research,and formed a model and method system of optimal allocation of land use,including mathematical programming method,simulation optimization model and intelligent optimization algorithm.However,there are still some deficiencies in the existing methods,which are mainly reflected in a series of urgent problems to be solved,such as the land use multi-objective decomposition method can not meet the expansion of land use multi-objective dimension,the optimization ability of intelligent algorithm needs to be further improved,the optimization algorithm does not fully consider the selection preference of decision-makers,and the optimization model can not meet the rapid response needs of actual decision-makers for intensive computing.Focusing on the two core problems of land resource quantity structure optimization and spatial layout optimization in land use optimal allocation,this paper proposes a multi-objective collaborative optimal allocation model of land use based on coupled objective decomposition evolutionary computing(MOEA / D)and swarm intelligence(SI).At the same time,the preference of decision-making subjects and land use domain knowledge are integrated into the model,and the intelligent algorithm is parallelized by cloud computing MapReduce programming mode.The paper finally formed the following research conclusions:(1)The multi-objective collaborative index system of land use composed of five dimensions of economy,society,ecology,environment and space is reconstructed.Using the function structure element method of system theory,the multi-objective collaborative index system and evaluation model of land use are established.(2)A multi-objective coordination mechanism of land use based on decomposition strategy is established.Aiming at the multi-objective system of land use optimal allocation,this paper adopts the multi-objective decomposition strategy based on Tchebycheff method to form a multi-objective collaborative optimization method of land use based on objective decomposition.The goal decomposition strategy expands the multi-objective optimization dimension of land use and improves the ability of the model to deal with multi-objective collaborative optimization.(3)A hybrid intelligent algorithm model of land use multi-objective optimization coupled with evolutionary computation and swarm intelligence is established.Based on the in-depth analysis of classical genetic algorithm and ant colony algorithm,combined with land use multi-objective collaborative system and multi-objective collaborative method based on objective decomposition,this paper forms a hybrid intelligent algorithm model of land use multi-objective collaborative optimization based on objective decomposition,that is,a land use optimal allocation model based on MOEA/D-ACO.(4)The coupling method of decision-making subject preference,domain knowledge and intelligent optimization model based on weight iteration method is established.In this paper,the preference multi-objective decomposition method based on weight iteration and land use domain knowledge are integrated into the intelligent optimization algorithm to form an intelligent land use optimal allocation model combining the preference of decision-makers and domain knowledge,which couples the configuration model of domain knowledge and decision-making preference of decision-makers,so that the optimization results converge to the expectation of decision-makers.(5)A parallelization method of intelligent optimization model based on cloud computing MapReduce model is established.In this paper,MapReduce parallel programming model based on cloud computing is adopted to realize the parallelization of knowledge mining algorithm in land use field and intelligent algorithm for multi-objective optimization of land use,so as to improve the scalability and availability of the model.(6)An empirical case of the model.The study selects the urban agglomeration in Central Yunnan Province as the empirical case object of the model.The results show that: ○1 the land use multi-objective collaborative system designed in this paper can effectively serve the land use optimization model and intelligent algorithm,and the evaluation results are close to the expectation of regional planning.○2 the optimal land use allocation algorithm based on decomposition mechanism is better than the current typical optimization algorithm based on Pareto mechanism in terms of multi-objective optimization ability.○3 the time complexity and space complexity of the hybrid intelligent algorithm based on domain knowledge guidance and decision-making agent preference decrease significantly in the solution process.○4 when dealing with intensive computing problems,the intelligent algorithm using cloud computing programming model can obtain better speedup ratio and parallel performance,and show better scalability and potential availability.Finally,this study has formed the following three innovations:(1)A multi-objective coordination method of land use based on Chebyshev method and weight iteration method is established.The Chebyshev method is used to realize the decomposition of land use multi-objective,and then the weight preference of each sub objective is automatically calculated by the weight iteration method,so as to realize the coordination between the preference of decision-making subject and the multi-objective decomposition mechanism,and finally provide basic support for the hybrid intelligent algorithm and model of land use multi-objective collaborative optimization in this paper.(2)A multi-objective collaborative optimization model of land use based on hybrid intelligent algorithm and decision-making agent preference is proposed.In order to give full play to the advantages of each kind of intelligent algorithm,this paper attempts to establish a land use multi-objective collaborative hybrid intelligent optimization model coupling MOEA /D algorithm and ACO algorithm,and explore the mechanism of coupling the preference information and domain knowledge of land use decision-makers with the intelligent optimization model.Finally,the organic connection and unity of land use multi-objective collaborative system,decision-making subject preference,domain knowledge,intelligent algorithm and land use optimization model are realized.(3)A parallelization method of land use related intelligent algorithms based on MapReduce model is proposed.Firstly,aiming at the problem of knowledge mining in the field of land use,this paper proposes a land use rule mining algorithm based on MapReduce coupled ant colony intelligence;Secondly,aiming at the multi-objective collaborative optimization algorithm of land use,the parallel algorithm of multi-objective collaborative optimization of land use based on MapReduce is introduced and designed to realize the cooperation between cloud computing platform,intelligent optimization model of land use and GIS. |
PDF Full Text Request