适应气候变化的分析框架及政策涵义

适应气候变化与发展议题密切相关,这也使得适应成为一个过于宽泛的命题,影响了对适应内涵的理解。首先,本文基于国内外对适应问题的探讨,提出了适应气候变化的基本分析框架,即基于不同发展阶段的适应需求,区分增量型适应和发展型适应,并通过工程性适应、技术性适应和制度性适应三种适应手段增强适应能力。增量型适应是指在系统现有基础上考虑新增风险所需的增量投入,这种适应所针对的是发展需求基本得到满足,仅仅需要应对新增的气候风险所需的适应活动;发展型适应是指由于发展水平滞后,使得系统应对常规风险的能力和投入不足,需要协同考虑发展需求及新增的气候风险。其次,分析了适应的四个基本步骤,即:①评估气候风险及脆弱性;②甄别各种可能的适应对策;③选择可行的适应措施;④推荐"成功"的适应行动。还介绍了相应的社会经济分析方法,如脆弱性评估、社会经济影响评估、适应的成本效益分析等。最后,基于中国适应气候变化的基本需求及优先领域,提出了相应的政策建议,如开发农业适应技术、加强流域综合治理、开展健康风险监测、实施灾害保险计划等。

地方治理中的增长性公共政策:内在结构、治理后果与改革方向

发展主义曾长期主导地方政府的发展思路。在发展主义范式下,经济发展成为地方政府行为的主要目标。经济发展与地方政府官员的行为逻辑相结合催生出地方治理中的代理型发展主义形态与增长型公共政策工具,即通过投资和土地两种政策驱动地方经济发展。增长型公共政策强调经济发展而忽视社会发展的重要性,因此需要从价值和结构等层面对其进行调整和完善。对增长型公共政策的调整实际上也引发了有关地方治理转型的改革议程。

Artificial Intelligence,Smart Robots and a New Economic Order

In the process of transition from agricultural society to industrial society,which started with the Industrial Revolution in England,the mechanization process experienced five different stages and in the last stage,with the development of computers,automation in production was achieved.While developments in a certain region or country of the world spread to other parts of the world with technological spread,technological revolutions also spread and paradigm changes occurred.With the development of information processing technologies,productivity has started to increase with the use of automation and robot technology in production.This process,which continued until the 2010s,is thought to lead to the formation of smart factories that can produce under the dominance of robots,after the new point reached in artificial intelligence and robot technology,and this development will further increase productivity in production.Intelligent robots working in the internet of things system will be able to work with greater power and longer periods than humans,and smart factories that are almost never shut down will emerge.In the transformation in this process,which is also called robonomics,changes in the theory of economy may occur and a new economic order may emerge.The question of why behind-the-scenes countries,such as Turkey,could not catch up with the leading ones,is another matter of discussion.However,in such periods of technological paradigm change,an opportunity arises for lagging countries for their economic development.On the other hand,we can say that Turkey will either be able to catch up with the technological level of developed countries by taking advantage of the opportunity,by means of a step-by-step technological development,or it will continue to stay among the countries that lag behind by missing the opportunity.

Contested incrementalism: Elemental’s Quinta Monroy settlement fifteen years on

Quinta Monroy is an award-winning co-designed settlement for 93 families on half a hectare of land at Iquique in northern Chile.Neighbors,complaints about the disorderly settlement peaked after the landowner^death and provoked untenured residents to seek government subsidies to redevelop the settlement.From 2003,a government social housing project was coordinated by the,,Elementar,architecture firm with US$10,000 per household.With the residenfs temporary relocation,93 modular and interlinked apartments were built around a series of courtyards.These apartments,which were designed as,fhalf-houses,M were subsequently co-opted by residents adding rooms in locations planned in advance by Elemental.Many households have since doubled the size of their apartment and reformed the settlement in ways not anticipated by Elemental.This paper details a spatial and ethnographic study of the Quinta Monroy settlement since redevelopment to identify opportunities and risks that accompany this type of social housing model.The study reveals evidence that residents,capacities to enlarge apartments commonly exceeds the architects expectations and that unregulated expansions often compromise the settlements livability.This research anticipates further opportunities for expansion in this semi-regulated settlement and investigates possibilities that another contested slum settlement may emerge.

基于增量式发育深度强化学习的无人机路径规划

为了克服深度强化学习训练时间长、收敛速度慢的问题,针对密集动态障碍环境下的无人机(UAV)路径规划,引入了增量式发育知识库,对深度确定性策略梯度(Deep Deterministic Policy Gradient,DDPG)算法进行改进。首先,根据模糊匹配的思想建立威胁模式知识库,将飞行过程中遇到的密集动态障碍作为知识进行增量式存储,避免对相似障碍环境的重复训练。其次,在底层DDPG算法规划安全航路的基础上构建避障策略知识库,根据威胁模式直接输出避障策略,缩短训练时间。最后,搭建发育式的威胁-避障映射关系,实现“线上实时避障,线下自主寻优”,不断提升UAV避障性能。对比实验结果表明,所提方法能极大提高DDPG训练效率,满足UAV在密集动态障碍环境中实时避障的需求。

发展型机器人实时特征提取方法研究

针对发展型机器人自主学习过程中特征提取涉及的增量计算和实时性问题,结合已有的CCIPCA(直观无协方差增量主成分分析)和BDPCA(双向主成分分析)算法,提出了一种增量的BDPCA算法.采用了迭代的计算方式,具备增量的计算能力,并且将2维原始图像矩阵直接作为处理对象,有效地降低了计算量,缩短了程序运行时间.以机械臂待抓取的物块作为实验样本,利用支持向量机进行分类,验证上述算法.实验结果证明了该算法的有效性,平均分类率可达90%,算法处理速度大约26帧/秒,基本满足了发展型机器人的实时处理需求.