Ali Mansourian
Professor
A comparing study between AHP, AHP-OWA and Fuzzy AHP-OWA multi-criteria decision making methods for site selection of residential complexes in Tabriz-Iran
Author
Summary, in English
A decision is the result of comparing one or more alternatives with respect to one or more criteria that we consider relevant to the decision. Decision making can be considered as one of the most important challenges that analysts and experts encounter to solve complex problems. In recent decades, different methods and algorithms have been presented to support decision making. Multi-criteria analysis (MCA) techniques are well-known decision support tools for dealing with complex decision constellations where technological, economical, ecological and social aspects have to be covered. To solve spatial problems, multi-criteria decision analysis is integrated with GIS, in which spatial data (maps) and evaluation values (analysts' priorities and criteria) are combined. Site selection is one of widely used spatial decision making problems. The purpose of site selection is to find a set of appropriate spatial alternatives for a particular application. Since site selection is influenced by a variety of factors, it is a multi-criteria decision making problem. Materials and Methods There are different approaches for GIS-based multi-criteria decision making. This paper intends to compare and evaluate three GIS-based multi-criteria decision making analysis including AHP, AHP-OWA and Fuzzy AHP-OWA for a site selection problem. Site selection of residential complexes in Tabriz, with emphasize on environmental factors, is the case study of this research. Cost, environmental pollution and land suitability were considered as three main factors, each of which included some criteria, and ten data layers corresponding to these criteria were collected and analyzed for the study. Discussion of Results First, an AHP method was used to determine quantitative information about the relative importance of the factors and criteria. In this regard, the problem was decomposed in a hierarchical structure. Then, the criteria/factors were compared to each other in a pairwise comparison matrix. After that, numerical values expressing a judgment of the relative importance (or preference) of one criteria against another were determined and assigned to each criteria. Finally, the weighted data layers were integrated and a suitability map for the construction of residential complexes was generated. As a result of this experience, it was highlighted that the three main factors are in different levels of risk and trade-off. For example, the "cost" criteria have a high compensation and moderate risk. On the other hand the "environmental pollution" criteria, due to the dangerous effects on human health, have high risk and low compensation. Factors relevant to "land suitability" are of low risk and low compensation. Since AHP cannot control the risk and trade-off issues in a multi-criteria decision making model, OWA approach was used. OWA is a family of multi-criteria combination procedures. It involves two sets of weights: the weights of relative criterion importance and the order weights. By specifying an appropriate set of the OWA weights, one can generate a wide range of different suitability maps. However, since OWA could not properly handle the experts' preferences, its integration with AHP was used for the site selection problem. With these in mind, at the second stage, the expert's preferences and ordered weights were determined and entered into IDRISI environment and then the suitability map for the construction of residential complexes was generated (Fig.1). This experience showed some defects of AHP-OWA approach for the site selection problem, as well. For example, in some problems with a lot of criteria, the exact detection of relations between the criteria could be very difficult. The AHP, OWA or combined AHP-OWA, can not satisfy the entry of qualitative information to decision making model. The integration of linguistic fuzzy quantifiers with AHP-OWA can be a solution to this issue. "Figure Present" Based on the type of linguistically quantified statements one can distinguish between: the absolute linguistic quantifiers and the relative (or proportional) linguistic quantifiers. Statements such as at least about 4, about 5, almost 10, not much more than 10 and more than 5 provide examples of the absolute quantifiers. The relative linguistic quantifiers indicate a proportional quantity such as most, many, a few, almost all, about half ana about 60%. They can be represented as fuzzy sets of the unit interval [0, 1]. They measure a proportion of a set, where 0 means 0% and 1 means 100%. At the third stage of the study, an integrated fuzzy AHP-OWA approach was adapted for the site selection problem. The model was decomposed into hierarchy structure model. Then the criteria and ordered weights were entered to the model by pair-wise comparison matrix and quantifier-guided ordered weights. Finally, the weighted data layers were integrated and a suitability map based on quantitative and qualitative information as well as controlling the risk and trade-off in decision making was produced (Fig.2). Conclusion The study and comparison of the maps resulting from the implementation of three multi criteria decision making methods to select suitable areas for residential complex construction in Tabriz city showed that in the cases where operators with high ORness (risk) values were applied, map cells had a higher average value than maps with less ORness. Studies also showed that, in high risk maps, more areas have been introduced as suitable areas for residential complex construction. On the other hand, it can be mentioned that the variety domain of cell values and the extent of the most suitable areas in AHP, OWA and Half Fuzzy linguistic quantifier were very close to each other (Fig.3). "Figure Present".
Publishing year
2011-05-01
Language
English
Pages
77-92
Publication/Series
Journal of Environmental Studies
Volume
37
Issue
57
Document type
Journal article
Publisher
Dānishgāh-i Tihrān, mu̓assisah-i muṭāli’āt-i muḥīṭ-i zīst
Topic
- Other Computer and Information Science
Keywords
- AHP
- Environment
- Fuzzy
- GIS
- OWA
- Artificial Intelligence (AI)
- Geospatial Artificial Intelligence (GeoAI)
Status
Published
ISBN/ISSN/Other
- ISSN: 1025-8620