The Research For Regional Sustainable Development Based On System Dynamics, Artificial Neural Networks And Computer Simulation

In this dissertation, sustainable development system was studied for Jianhu County,Jiangsu province, resorting to system dynamics, artificial neural networks, by syncretizingthese methods to simulate system future trends in variant development modes,simultaneously, the appraisal model was established, and so the best future developmentroute of region could be selected. Because the System with property of high intricated andnon-experimental, the accurated mathematical model can't be established, so the researchobjective is to explore a new way to create systematic model for sophisticated system,i.e,using artificial neural networks to simulate the relationship which can't be ensured easily.On the basis of systematic model, the system dynamic response was simulated in variantparameters; it is significance for us to kown and manipulates sustainable developmentsystem.In this dissertation, the historical status of regional sustainable development was firstlyinvestigated and studied elaborately, and gleaned abundant system historical data, based onthese information, the mathematical modeling for regional development system was created,the sustainable development system was analyzed and depicted by theory of sustainabledevelopment and research objectives, and the decomposed the system as three subsystem,i.e., population subsystem, economical subsystem, resource and environment subsystem,they are tied up closly and interacted compactly, then 95 systematic variables was appliedto describe these subsystems and whole system; Secondly, the system dynamics syntax wasused to describe the systems, and classified 95 systematic variables into three kinds, i.e.,state variables, velocity variables and auxiliary variables. Based on analysis of innerstructure and cause-effect feedback loops in these systems, make sure every subsystemintramural construction and primary feedback loop and secondary feedback loop. Based onthese feedback loops, state equation, velocity equation and auxiliary equation wereestablished, therefore, every subsystem modeling could be confirmed, then three subsystems were sophisticatedly multi-dimensionally coupled into whole regionalsustainable development system of Jianhu County. In the system model, most velocityequations and auxiliary equations were established by artificial neural networks, usingnetworks learned the historical data, using system recognition to found their relationships,the relationship founded by using such methods can easily pass through historical test, andmake system modeling have behavior similarity with real system, and pass throughbehavior consistency validation, so system modeling can simulate real system developingstatus in the future. Because it needs plenty of historical data to establish artificial neuralnetworks modeling, the historical data is usually not enough, so 3-order B spline functionwas applied to construct learning samples by interpolating historical data. The artificialneural networks modeling established on the basis of enough learning and training sampleshave high robust and generalized capacity, it can predict robustly the future tendency ofsystem.After establishing the system modeling and pass it though behavior-similarity validationtest, sensitivity analysis for system modeling was carried out, doing so is to designoperating parameters, and based on these parameters' different values, variant developmentmodes of real system could be simulated. System sensitivity analysis includes three types:value, behavior and policy sensitivity.The development modes in 2001-2020 of Jianhu County sustainable development systemwas reduced to 4 kinds on the basis of operating policy parameters variant values, they are:Traditional Development Scheme, High Economical Scheme, Resource and EnvironmentProtected Scheme, Economy-Environment Coordinate Scheme.The Traditional Development Scheme means system development was little imposed on,it'll suit itself to the future, system parameters have little change. From the simulatiingresult, the people amount increasing rapidly, the velocity of economical development isrelatively slow, resource and environment face to severe damage, in the long term, and it'llresult in disadvantaged effect to local economical development, social progress andenvironment protected. So this development mode must be discarded in the future.The High Economical Scheme means keep economy preferential development on thepremise of sustainable development. The parameters in primary feedback loop ofeconomical subsystem was changed in order to be propitious to economical development,the parameters in other subsystems was changed little, consequently form High EconomicalScheme. From simulating result, under his mode, economy will develop rapidly, social enterprises will progress expeditiously, and people's living conditions will improve highly,but after many years stretch, the sustainability of regional development will be badlyinfluential, regional development isn't under high sustainable level, and result in seriesquestion;Resource and Environment Protected Scheme emphasized on protecting local resourceand environment, as same as High Economical Scheme, some parameters in resource andenvironment subsystem feedback loop have changed, from simulating result, under thisdevelopment modes, local resource and environment were getted effective protection, andsimultaneously economical development and social progress is remarkable. But generallyspeaking, because it lays particular stress on resource and environment protection, so it willrestrain economical development in the end, under this mode, the indicators valuecorrelative with economy raise slowly, so it'll lead to economical development obstructedin the long run. The problems brought by economical development lag will severely harmfor regional sustainable development;Economy and Environment Protection Scheme means give attention to both economicaldevelopment and resource and environment protection. Although its economical increasingspeed is less than High Economical Scheme and its resource and environment protectioneffect are lower than Resource and Environment Scheme, at first sight, it is not idealscheme, this scheme makes economical development harmonize with resource andenvironment protection exactly. Under this mode, from simulating result, economy getrapid development, resource and environment get effective protection; it makes localeconomical benefit, social benefit and ecological benefit get full satisfaction, everyindicators can up to scratch, i.e. it contents with criteria of check and accept constituted byNational Environment Protection Chief Bureau, it achieves objective based on highsustainability, so it'll benefit for local perennial and steady-going sustainable development.As same as establishing system model, It is very tough job to appraise regionalsustainable development level, because there is very complicated relationship among theevaluating indicators, so traditional methods unfitted with disposing of such problem.Using traditional methods will induce more subjective as establishing evaluating model,and model is easy to be molested by basic data, so Learning Vector Quantization Networkswas applied to establish evaluating model for regional sustainable development level, onthe basis of system modeling and simulation, the sustainable development levels of JianhuCounty in every past years were appraised synthetically according to 36 evaluating indicators. The appraisal for local sustainable development level from 1990 to 2000 wasfirstly carried out, and then appraised the level in the future years from 2001 to 2020 under4 types' development modes on the basis of simulating result, and based on the appraisal,the best development mode for future can be selected.From appraising result, the sustainable development level in Traditional DevelopmentScheme mode descend fast, in the 20 years period, system sustainable level always keep onthe lower grade, and conclusion can be drawn that the system will run into non-sustainablestatus before long;The level in High Economical Scheme mode kept on high grade on the beginning, butsome years went by, till 2013 year, system sustainable development level commencedescending, and attain weak sustainable level at the end of simulation;The Resource-Environment Protected Scheme incarnates similar developing trend asHigh Economical Scheme;The Economy-Environment Coordinate Scheme makes economy, society and ecologicalenvironment harmonious development, it'll lead to system coming into good circulation, sothe system sustainable level always keep on the higher grade. It is assumed that, the systemmust progress forward on such high sustainable level. From appraising result, conclusioncan be drawn that Economy-Environment Coordinate Scheme is the best developing modefor Jianhu County, so the policy that development abode by Economy-EnvironmentCoordinate Scheme could be constituted according to the value range of parameters insystem modeling.The methodology applied in the dissertation is a new attempt to established systematicmodel for sophisticated systems, it proves that, as artificial intelligence was used increating systematic model, the time consuming and labor intension can be cut down largely,the system model with robust for data noise and itself run mode, and is easily understood,the simulating conclusion is real and authentic. No other methods can parallel it. The resultin the dissertation can provide scientific reference and guidance based on simulatingexperiment for regional sustainable development.