统计与数据科学学院知识库

摘要

从韧性理念的不断丰富到与复杂主体城市相结合，城市韧性理念与当代可持续发展理念相契合，成为引领城市发展且备受热议的话题。准确地进行城市韧性量化评估，探究体系间内在机理，寻求适宜的城市韧性提升策略是当前研究的关键。论文对甘肃省各市州城市韧性进行量化评估并探究子系统间相互作用程度，寻求适宜甘肃省城市韧性发展的提升策略。

首先，论文梳理相关文献，基于城市韧性内涵和外延，分别从经济系统、社会系统、基础设施系统和生态系统方面建立城市韧性指标体系。采用熵值法作为指标评价函数，对甘肃省城市韧性现状分别从省域、市域角度进行演化分析，引入耦合协调度模型，从省域、市域角度探索分析甘肃省各市州的经济、社会、基础设施和生态系统耦合协调情况及演变特征，分析城市韧性体系的内在机理。

其次，论文结合遗传算法、极限学习机构建一种基于遗传算法-极限学习机（GA-ELM）的城市韧性模拟方法，基于指标评价函数得到的城市韧性值的基础上，引入遗传算法对模型参数进行寻优，进一步提升ELM模型模拟精度，运用遗传算法优化的ELM模型，结合十四五规划目标对城市韧性进行动态仿真模拟，进一步提出针对甘肃省的城市韧性提升策略。

第三，论文首先通过分析甘肃省城市四大系统的耦合协调情况，与GA-ELM模型模拟的城市韧性提升策略相结合，总结甘肃省城市韧性现状，并尝试对甘肃省城市韧性情况提出思考和展望，为政府制定相应改善政策提供参考。

关键词：城市韧性 耦合协调度模型 GA-ELM神经网络

Abstract

From the continuous enrichment of the concept of resilience to the combination with the complex main city, the concept of urban resilience is consistent with the contemporary concept of sustainable development, and has become a leading and hotly discussed topic of urban development. The key of current research is to accurately quantify urban resilience, explore the internal mechanism between systems, and seek appropriate strategies for improving urban resilience. This thesis quantitatively evaluates the resilience of each city in Gansu Province and explores the degree of interaction between subsystems, so as to find the appropriate strategies to improve the resilience of Gansu province.

Firstly, based on the connotation and extension of urban resilience, an index system of urban resilience was established from the aspects of economic system, social system, infrastructure system and ecosystem system. The entropy method is adopted as the evaluation index function, the toughness of city, Gansu province, respectively from the perspective of provincial and regional evolution analysis, introducing the coupling coordination degree model, from the perspective of provincial and regional exploration analysis of Gansu province cities in the state of the economy, society, infrastructure and ecological system coupling coordination situation and evolution characteristics, the analysis of the internal mechanism of urban resilience system.

Secondly, the thesis combines the genetic algorithm and machine learning to build a limit - extreme learning machine based on genetic algorithm (GA) - ELM city, toughness simulation method based on the urban toughness values obtained from index evaluation function, on the basis of introducing the genetic algorithm to optimization of model parameters, further enhance the ELM model simulation precision, using genetic algorithm to optimize the ELM model, Combined with the goals of the 14th Five-Year plan, the dynamic simulation of urban resilience was carried out, and the strategies for improving urban resilience in Gansu province were further proposed.

Thirdly, by analyzing the coupling coordination of the four urban systems in Gansu Province and combining with the urban resilience improvement strategy simulated by GA-ELM model, the thesis summarizes the status quo of urban resilience in Gansu Province, and tries to put forward thoughts and prospects for the urban resilience improvement strategy in Gansu Province, so as to improve the cities in Gansu province for the government.

Key words: Urban resilience; Coupling coordination degree model; GA-ELM neural network