从物理模型到智能分析——降低天气预报不确定性的新探索

本文回顾了气象预报技术发展的进程,人们不断通过创造性的智慧发挥,从传统的经验判断到依靠现代数学物理方法,实现通过模式方程组求解大气演变过程,取得了成功。但面对复杂的天气气候系统,许多非线性不确定问题仍难以解决。随着人工智能(AI)技术的发展,有希望助上一臂之力。文中介绍了一些初步设想和进展,在观测识别、数据处理、短时临近预报、模式参数化、预报结果集成分析等环节都有初步成果,进一步的研发、拓展值得期待。     

“五维协同”下的测绘地理信息成果质量检查软件体系初探

质量检查软件是质量检查的重要手段,是影响质量检查工作的重要因素。本文在分析测绘地理信息成果质量检查软件的现状和存在问题的基础上,结合现有的手段和新的技术,提出了由成果维、项目维、大数据维、智能维、历史维组成的“五维协同”下的测绘地理信息成果质量检查软件体系,分别阐述了测绘地理信息成果质量检查在各个维度下的工作内容,介绍了5个维度在业务流程和数据流动之间的关系,为测绘地理信息成果质量检查软件之后的研究内容、发展方向提供了思路。     

多源大数据支持的土地储备智能决策模型集研究

伴随着城镇化的快速推进,城乡建设用地资源日益紧张,但目前土地储备决策缺乏精准科学依据,无法有效地进行资源配置和宏观调控。针对此问题,本文深入剖析土地储备基本业务与决策环节,研究了一套面向土地储备的智能决策模型集,包括存量土地监测模型、收储成本预测模型、出让价格预测模型、储备平衡分析模型、相似地块分析模型、开发时序分析模型及病态地块识别模型,旨在将土地储备决策环节科学化、定量化和模型化,并重点为土地储备总量、效益、规模、结构、布局、时序的统筹安排提供建议。另外,该模型集具有体系化、高效灵活、智能化的特点,能够服务于储备业务全链条,满足即时决策应用需求和实现模型自主更新与进化,保证模型的时效性。最后,该模型集已经工程化应用于宁波市土地储备智能决策支持平台,实践验证了以上决策模型具有较高的准确度和实用性,表明模型集能够为土地储备的科学决策提供理论依据,有利于土地资源的集约利用和高效配置。     

城镇化与生态环境耦合动态模拟模型研究进展

按照地理学科发展趋势,对城镇化与生态环境耦合的研究将由定量描述转入动态模拟。目前,城镇化与生态环境耦合动态模拟模型呈现多元化。论文系统梳理了其中4类常见的动态模拟模型,包括城镇化与生态环境耦合系统动力学模型、基于人工智能算法的城镇化与生态环境耦合动态模拟模型、基于土地利用变化的城镇化与生态环境耦合动态模拟模型以及基于多模型集成的城镇化与生态环境耦合复合模型。主要结论如下：系统动力学模型被广泛应用于城市复杂系统、城市转型和可持续发展以及城镇化与生态环境单要素耦合的动态模拟之中,但存在空间解释不足以及忽视系统自适应性等问题;人工智能算法(ANN和BN)在自学习、自组织、自适应系统或不确定性系统模拟中具有显著优势,并被应用于城市扩张、环境变化、资源需求以及生态脆弱性的识别之中,但应用面相对狭窄且限制条件偏多;土地利用变化模型(CLUE/CLUE-S、CA和MAS)局限于从土地城镇化视角模拟城镇化与生态环境耦合;基于多模型集成的复合模型实现了各模型之间的优势互补,已成为城镇化与生态环境耦合动态模拟模型的发展趋势。今后,应从技术和理论2个层面实现城镇化与生态环境耦合动态模拟模型的进一步发展,并加强对微观过程的模拟。     

从徐志摩和顾城的爱情诗透视其性格与命运

“新月”诗人徐志摩和“朦胧”诗人顾城是二十世纪中国文学史上的两位天才诗人。爱情无疑使他们的创作充满了灵性,但爱情也使他们走上了不归路。徐志摩“魄散天外”;顾城“魂断异国”。爱情激发了诗人生命中的灵性,爱情也摧毁了诗人充满灵性的生命,这大概就是诗人难以逃脱的命运。     

Features of geonomic prognostic functions for the maximum possible earthquake

The recent artificial intelligence techniques are commonly applied to solving problems in which multidimensional statistical analyses of various quantities and their modelling prognostic functions can be used. This paper attempts to summarize the characteristics of the prognostic functions applied in the determination of the maximum possible earthquake. Geonomic quantities are reviewed and categorized with respect to their influence upon the estimation of the maximum possible earthquake.     

A bat-inspired approach to define transition rules for a cellular automaton model used to simulate urban expansion

A new metaheuristic approach is presented to discover transition rules for a cellular automaton (CA) model using a novel bat movement algorithm (BA). CA is capable of simulating the evolution of complex geographical phenomena, and transition rules lie at the core of these models. An intelligence algorithm based on the echolocation behavior of bats is used to discover explicit transition rules for use in simulating urban expansion. CA transition rules are formed by links between attribute constraint items and classification items. The transition rules are derived using the BA to optimize the lower and upper threshold values for each attribute. The BA-CA model is then constructed for the simulation of urban expansion observed for Nanjing City, China. The total accuracy of newly formulated BA-CA model for this application is 86.9%, and the kappa coefficient is 0.736, which strongly suggest that the interactions of bats are effective in capturing the relationships between spatial variables and urban dynamics. It is further demonstrated that this bat-inspired BA-CA model performs better than the null model, the particle swarm optimization-based CA model (PSO-CA), and the ant colony optimization-based CA model (ACO-CA) using the same dataset. The model validation and comparison illustrate the novel capability of BA for discovering transition rules of CA during the simulation of urban expansion and potentially for other geographic phenomena.