应对人口流失的景观设计改进方法:以美国宾夕法尼亚州约翰斯敦的 城市更新为例

Over the last 50 years, 370 large cities worldwide have severely depopulated, or shrunk, by at least 10%. Johnstown, Pennsylvania, is the third fastest U.S. shrinking city. Primarily a victim of deindustrialization, Johnstown faces severe decline issues related to depopulation, including social disorder and lowered quality of life. This project develops a framework for urban design for shrinking cities, integrating permanent functions into high development potential areas but temporary functions into declining areas. This approach allows for future development to occur through time as the city recovers. Using a GIS-based weighted overlay model to assess the threat level of decline, 4 sites were identified and strategies for each were developed. Master plans to retrofit new functions integrating residents' desire and demands into vacant / abandoned properties were then generated for each site. Rather than chasing hefty attempted quick-hitting developmental incentives, this approach will bring new long-term economic engines and lifestyles to the city due to a diversity in the economic base; it also pays attention to the social dimension of urban regeneration by providing a structured way to promoting social justice and equity in shrinking cities.     

路缘沉积:利用原型思维过程探究如何通过路缘及路肩设计截留暴雨中的路面污染物

The design of a curb is straight forward. The curb itself provides a conveyance of stormwater, facilitating the movement of water and pollutants from the street into waterways. Pollutants such as sediment, nutrients from lawn fertilizers, bacteria, viruses, pesticides, metals, and petroleum byproducts accumulate on the road surface and are released during storm events, carried to storm drains, and deposited into waterways, often without treatment. Once pollutants enter the waterways they impact the ecosystem and affect water quality. How can discrete standards—like a curb—be leveraged to have larger systemic impacts? The redesign of the curb to perform as a magnet for pollutants can challenge this design standard. During the summer of 2019, the interdisciplinary research team tested alternatives to the standard concrete curb and apron at Ohio State University. The team used an iterative design process to add patterning and crenellations to the face of the curb and apron. Using full scale models to test simulated storm events, the team collected data to evaluate the performance of 21 alternative designs. The results suggest the new combined curb and apron designs can abstract pollutants from roadways before they are detrimental to water bodies and aquatic ecosystems.     

一种并行化的分类算法研究

研究并实现了基于Spark的KNN算法的并行构建.分析了MapReduce模型和Spark在处理迭代计算方面的优劣,结合KNN算法的自身特点设计了对应的Map算子和Reduce算子,实现了KNN算法的Spark并行化.实验结果表明,较传统的KNN串行算法和MapReduce并行KNN算法,基于Spark的并行KNN分类算法具有较好的效率和较高的可扩展性.