DSpace Repository

Hybrid Multi-Criteria Evolutionary Algorithms for Optimization Problems in Sustainable Landuse Planning

Show simple item record

dc.contributor.advisor Tripathi, Nitin Kumar
dc.contributor.author Handayanto, Rahmadya Trias
dc.contributor.other Guha, Sumanta
dc.contributor.other Kim, Sohee Minsun
dc.date.accessioned 2018-01-19T01:40:29Z
dc.date.available 2018-01-19T01:40:29Z
dc.date.issued 2018-01
dc.identifier.other AIT
dc.identifier.uri http://www.cs.ait.ac.th/xmlui/handle/123456789/882
dc.description.abstract Cities worldwide have been using land use zoning to handle their unprecedentedly rapid urban-growth. Land use zones are widely proposed based on the sustainable urban form. This study used two urban form, namely the compact city and the eco city, as urban form foundation. Based on these urban forms, an optimization module was created for land use in Bekasi City, Indonesia. Four criteria functions are derived from compact city and eco city, i.e. compatibility, dependency, compactness, and suitability. These criteria functions were optimized through hybrid optimization method which combined two evolutionary algorithms, i.e. particle swarm optimization and genetic algorithm, and a local search method (pattern search). The proposed optimization-method was used to create the three zones (commercial, industrial, and residential) after optimizing the current land use (2015). It also used to allocate expected new land uses (2030) with three scenarios, i.e. business-as-usual, sustainable development, and government policy. Future land uses are predicted through a nonlinear autoregressive neural network with external input. Optimization result showed that the sustainable development scenario had better performance than business-as-usual and government policy scenarios. As a complement of land use optimization, the second part of the current study proposed the land use and land cover change models. Firstly, the types of urban growth within the study area were identified through a combination approach of remote sensing, GIS, and spatial metrics analysis. Characterizing growth types found that infilling and edge-expansion were dominant in Bekasi city. These growth types are feasible to be used as driving factors of land change modelling. A multilayer-perceptron neural network was used for predicting the urbanization (with and without urban growth consideration). It was found that a scenario with urban growth types as driving factors were more accurate than others. Prediction maps for 2030 and 2050 were also produced through this approach with two approaches including conservation and business-as-usual scenarios. Simulation result showed that conservation scenario could minimize the effects of diminishing vegetation, bare land, and agriculture. The main contribution of this research is the use of sustainable development concept in land use optimization. To answer the research problems, four objectives have been completed, including implementation of hybrid multi-criteria evolutionary algorithm in Bekasi city, comparing the sustainable-development based optimization with business-as-usual based optimization, comparing the government-policy based optimization with sustainable-development based optimization and business-as-usual based optimization, and implementation of the urban-growth prediction through another vertical application as a complement to land-use optimization. This study shows that the proposed method was able to optimize land use, compatible with another vertical application, and adding more comprehensive analysis to solve the problems within the study area. en_US
dc.description.sponsorship Ministry of Research, Technology and Higher Education (DIKTI) Indonesia en_US
dc.language.iso en en_US
dc.publisher AIT en_US
dc.subject evolutionary algorithm en_US
dc.subject hybrid optimization en_US
dc.subject multi-criteria analysis en_US
dc.subject compact city en_US
dc.subject eco city en_US
dc.subject land use zoning en_US
dc.subject driving factors en_US
dc.subject land use and land cover change en_US
dc.subject sprawl en_US
dc.subject neural network en_US
dc.subject urban growth modelling en_US
dc.subject spatial metrics en_US
dc.title Hybrid Multi-Criteria Evolutionary Algorithms for Optimization Problems in Sustainable Landuse Planning en_US
dc.type Dissertation en_US

Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace

Advanced Search


My Account