orange-fritters / urban-design

[CEE302 Final project] Resolving MCLP Using Walking Graph Network and Genetic Algorithm for the Proposal of Elder Facilities Locations in Gwanak-Gu

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Title

Resolving MCLP Using Walking Graph Network and Genetic Algorithm for the Proposal of Elder Facilities in Gwanak

The final submission for 2022-F CEE Urban Design Final Project

Abstract

While Gwanak-gu has the fifth-highest percentage of elderly residents living alone in districts of Seoul, there are few elderly leisure facilities, therefore polarization is extreme. In this project, a walking graph network and genetic algorithm were used to offer new locations for elderly leisure facilities in Gwanak-gu. This study suggests a novel approach to calculate the score of the distributed locations using graph theories to solve MCLP problem.

Data

This project used the data from map.seoul.go.kr

  • Distribution of Elder Populations
  • Facility Supply Map
  • Facility Demand Map

Graph Embedding Using QGIS

This project used QGIS to obtain coordinates from the tif images and combined them together

Formula

Suggested Formula is below.

The objective function had to be created since there was no appropriate reference for the objective function that was to be used with the genetic algorithm. With a graph network, this formula was suitable as an objective function since it could perform complicated computations of the 4390X4390-sized matrix, result in quantitative values, and reflect the characteristics of the adjacent matrix. S is the importance score, L is the NX1 matrix of the location candidate obtained from the genetic algorithm's process, and D is the physical distance that cannot be represented on the node. These scores for each node are weighted to obtain the score of the location selection itself. One scalar-type score is produced on the graph by a matrix product.

Result

The result is below.

Reference

[1] 김다솜, "서울 독거노인 35만명 넘어, 10명 중 3명 “고독사 우려" 「연합뉴스」, 2022년 5월 8일. 
[2] 정다은, 정광진, 유석연.(2021).서울시 지역생활서비스시설 불균등 분석.한국도시설계학회지 도시설계,22(2),59-78.
[3] 서울열린데이터광장(보건복지부 「노인복지시설 현황」), 2022.10.17
[4] 이태수, "노인의 생활반경은 ‘도보15분’ 경로당 자주, 오래 방문”  「연합뉴스」, 2017년 10월 12일
[5] “노인여가복지 인프라구축”, 서울시 정책지도, 2022년 10월 16일 접속, https://map.seoul.go.kr/spm/gly/policy/view.do?POLICY_CL_CODE=PLCAT20000&POLICY_NO=105
[6] 서울열린데이터광장(서울시 스마트도시담당관 「서울특별시 자치구별 장래인구추계」), 2022.10.17
[7] “Estimation of portfolio/SIP return using Monte Carlo-Python”, Medium, 2022년 10월 16일 접속, https://medium.com/@dstrader/monte-carlo-simulation-in-finance-python-part-2-13eb7bd89220
[8] “Genetic Algorith”, Papers with Code, 2022년 10월 16일 접속, https://paperswithcode.com/method/ga
[9] 조성아, 이건학. (2017). 공간 통계를 활용한 서울시 노년 인구 거주지와 노인 수요 시설 분포의 공간적 불일치 탐색. 한국도시지리학회지, 20(2), 99-112.
[10] 원종준, & 안건혁. (2010). 공공도서관 입지 및 시설특성이 이용활성화에 미치는 영향 연구: 서울시 공공도서관을 대상으로. 대한건축학회 논문집-계획계, 26(2), 79-86.
[11] 윤정미, & 이신훈. (2010). 효율성과 형평성을 고려한 공공시설 입지분석에 관한 연구-금산군 문화시설을 대상으로. 한국지리정보학회지, 13(2), 1-10.
[12] 마세인, & 김흥순. (2011). GIS 네트워크 분석을 활용한 노인복지시설의 접근성 연구: 인천시 내륙부를 중심으로. 국토연구, 61-75.

About

[CEE302 Final project] Resolving MCLP Using Walking Graph Network and Genetic Algorithm for the Proposal of Elder Facilities Locations in Gwanak-Gu


Languages

Language:Jupyter Notebook 100.0%