Using GeoTools to Implement the Multi -Criteria Evaluation Analysis : Weighted Linear Combination Model

Sammanfattning: Multi-Criteria Evaluation Analysis (MCEA) is a decision making method employed to solve optimization problems that require the evaluation of a set of criteria to build an optimal decision. Where MCEA is integrated with GIS, optimization problems with spatial aspects, such as finding the best location of a facility across an area, could be solved. MCEA is implemented through different criteria decision rules among of which is the Weighted Linear Combination (WLC) which has five different phases: 1) criteria definition, 2) criteria normalization, 3) weights definition, 4) criteria weighting, and 5) weighted criteria sum-overlaying. Free and Open Source GIS Software (FOSS GIS) is getting more popular. Indicators on this popularity include: 1) the number of governmentally funded projects, 2) the number of downloads, 3) and the increasing number of use cases of open source GIS software. Some governments have even decided to replace their proprietary software with FOSS software. Implementation of MCEA models in GIS software varied among loose-coupling, tight-coupling, and full-integration. In loose-coupling, GIS and MCEA are coupled through a file exchange mechanism. In tight-coupling, both systems share a common interface. In full-integration, custom programming scripts could be written to implement different MCEA tasks and models. Previous studies showed that most GIS software have limited capabilities to perform MCEA except for IDRISI, Common GIS, ILWIS, and TNT-GIS. Among of them, only ILWIS is an open source software distributed under GPL license. In this thesis, a desktop computer application, MC-Analyst, has been developed using GeoTools, a Java open source GIS software library, Java SE 1.7, and NetBeans IDE 7.3 to implement MCEA-WLC model. The milestone works of the thesis included: 1) The feasibility of using FOSS in general and FOSS GIS in specific was researched, 2) A survey of the literature that researched previous efforts and implementations for integrating MCEA in GIS software was carried out, 3) Experiencing software development using a FOSS GIS library, GeoTools, in order to address strength and weakness points for development using GeoTools and FOSS GIS, and 4) Some recommendations concerning future development approaches to produce better integrated MCEA and FOSS GIS systems were addressed for future works. To test MC-Analyst, a previous case study that located best sites to build solar farms in Colorado was employed. The output of the case study application was examined by altering input criteria weights for the main layers of solar DNI, distance to transmission lines, and distance to primary roads in the GIS-MCEA/WLC model. Different outputs found as expected that when a higher weight had been given to one criterion, the final suitability result showed noticeable higher suitability in areas where that factor was considered optimal. Raster and vector data processing capabilities of GeoTools were explored to perform basic GIS functions and it showed strong support for such functions. It was also found that in order for MC-Analyst to work as a standalone fully integrated Multi-Criteria Spatial Decision Support System (MC-SDSS), it would require other tasks that are found in common desktop GIS applications such as re-projection, clipping, and resampling. Therefore, re-using the source code of an open source desktop GIS application such as uDig or gvSig and integrate the developed MCEA-WLC model with it to produce fully integrated MC-SDSS could be of great interest for future development for FOSS GIS community. This would create an effective product that might compete with effective proprietary GIS software that implements MCEA models. The study showed also that before choosing a software library to start with, the development team should be acquainted with other possible open source and free software applications and libraries in order for the development process to be smooth and to output a productive application.