Using Machine Learning in Decision-making to Augment Beauty, Resilience, and Sustainability Outcomes in Urban Planning

拟议的研究旨在回答机器学习如何用于决策的问题，以增强英国城市规划、设计和工程的美感、弹性和可持续性。到2050年，预计将有67亿或68%的世界人口居住在城市（联合国，2019年），这意味着发展中国家和发达国家都面临着巨大的挑战，必须在今天做出关键的战略决策。为了确保这种城市化现象和其他持久的全球气候变化大趋势不会导致这些城市的环境灾难或生活质量低下，全球各国政府和城市领导人正在努力应对“复杂环境中的复杂情况”的决策（Bennett和Bennett，2008）。通过第四次工业革命扩大的数字连接和数据分析能力的同步增长产生的大量数据为决策和政策制定提供了良好的支持系统，然而，许多政策或战略以及适用于城市、环境或可持续性领域的设计和工程实例仍然是专制决策的遗产，创造了无效的，标准化选项和清单的非政治化解决方案（Jordan and Turnpenny,2015）。城市继续以一种可能不可逆转的方式发展和变化，因此，承认城市建筑环境的挑战非常复杂，在气候变化、技术进步、恐怖威胁、网络攻击、生物战等背景下理解这种复杂性，理解寻求解决方案的问题（本身可能是一个迭代过程），并理解人类的局限性是重要的第一步。并将帮助确定需要在哪些地方和哪些类型的决策来提高城市的表现，以及技术可以在哪些地方和如何提供帮助。这些领域的文献是有限的。为了取得成功，麦肯锡的广泛建议是，城市领导人采取战略方法，规划变革，整合环境思维，并将城市的价值主张建立在为所有人提供机会的基础上。这些建议要求今天领导城市规划、设计和工程（UPDE）的利益相关者的决策不仅要具有创新性、协作性和创造性，而且要有能力在一个层面上从过去学习，在另一个层面上看到并塑造未来。机器学习——人工智能的一个子集——凭借其对趋势和模式进行分类的能力以及处理多维、多变量大数据的能力，处于有利地位：1。帮助提高对这种情况的复杂性的理解。2 .改变新城镇、城市和大型混合用途城市开发（LMUDs）的构思、开发和维护方式；增加新城镇、城市和低成本城市的绩效成果。

The proposed research seeks to answer the question of how machine learning may be used in decision-making to augment beauty, resilience, and sustainability outcomes in urban planning, design, and engineering in the UK. BackgroundBy 2050, 6.7 billion or 68% of the world's population is expected to live in cities (UN,2019), which means that both developing and developed worlds are faced with enormous challenges and critical and strategic decisions that must be made today. To ensure that this phenomenon of urbanisation and the other enduring global megatrend of climate change do not result in environmental catastrophes or a poor quality of life in these cities, governments and city leaders, globally, are grappling with decision-making for "complex situations in complex environments" (Bennett and Bennett,2008). Vast amounts of data generated through the amplified digital connectivity of the Fourth Industrial Revolution and concurrent growth in data analytics capabilities have permitted good support systems for decision and policy-making, yet, many instances of policy or strategy as well as design and engineering as applicable to the urban, environmental or sustainability domains remain the legacy of authoritarian decisions, creating ineffective, depoliticised solutions from standardized options and checklists (Jordan and Turnpenny,2015). Cities continue to evolve and change in a potentially irreversible manner, so, acknowledging that the challenges of the urban built environment are extraordinarily complex, understanding this complexity in the contexts of climate change, technological advancements, terror threats, cyberattacks, biological warfare, etc, understanding the problem(s) (which in itself may be an iterative process) for which solutions are sought, and understanding the human limitations are important first steps, and will help with identifying where and what types of decisions are required to augment the performance of cities, and where and how technology could help. Literature is limited in these areas. To be successful, McKinsey's broad recommendations are that city leaders adopt a strategic approach, plan for change, integrate environmental thinking, and base the value proposition of their cities on opportunities for all. These recommendations demand that decision-making by stakeholders leading urban planning, design, and engineering (UPDE) today is not only innovative, collaborative, and creative but also based on an ability to learn from the past at one level and see and shape the future at another. Machine Learning - a subset of Artificial Intelligence - with its well-documented ability to classify trends and patterns as well as ability to deal with multi-dimensional, multi-variable big data, is well-placed to: 1. help improve one's understanding of this situation's complexity 2. transform how new towns, cities, and large-scale mixed-use urban developments (LMUDs) are conceived, developed, and maintained, and 3. augment the performance outcomes of new towns, cities, and LMUDs.