Keywords: site selection, GIS, expert systems, multi-criteria decision making, AHP.

Capital improvement facilities are major, long-term investments for owners and investors. Site selection of a capital project is a critical decision that significantly affects profit and loss of investors. The goal in a site selection exercise is to find the best location with desired conditions that satisfy predetermined selection criteria. The process of selection could involve a large number of candidate sites. Site selection typically involves two main phases: (a) site screening (i.e., identification of a small number of candidate sites from a broad geographic area and a range of selection factors) and (b) site evaluation (i.e., in-depth examination of each candidate site to the most suitable) [1]. A number of tools have been used to determine the proper site for a capital improvement facility. These tools include geographic information systems (GIS), Expert Systems (ES), and Multi-criteria decision-making (MCDM) techniques. Although these tools have played an important role in solving site selection problems, each tool has its own limitations.

Geographic information systems (GIS), Expert Systems (ES) techniques and Multi-criteria decision-making (MCDM) have been used in solving site selection problems for the last three decades. However, each of these techniques has its own limitations in dealing with spatial decision problems such as site selection problem. In the following sections, a brief introduction to ES, GIS and MCDM and their limitations in dealing with the sitting problem is provided.

GIS have often been used to identify suitable areas for land developments and the use of GIS in sitting analysis started in the late 1970s [2,3]. The success of GIS in sitting problems was attributed to its ability to perform deterministic overlay and buffer operations. GIS however, while possessing ideal capabilities for performing spatial searches based on nominally mapped criteria, are of limited use when multiple criteria and conflicting objectives are considered in the analysis.

Expert systems (ES) have achieved considerable success in many fields in recent years. Expert systems, however, lack the necessary mechanism to derive solutions based on spatial knowledge of different sites.

Multi-criteria decision-making (MCDM) Techniques were designed to analyze decision problems, generate useful alternative solutions, and to evaluate the alternatives based on a decision maker's values and preferences. These techniques, however, assume homogeneity within the study area, which is unrealistic in many spatial decision making situations such as site selection problems.

This paper presents a new approach in which these three tools were combined in a manner so that the shortcoming of one tool is complemented by the strength of another. In the proposed approach, an expert system is used to: (a) assist the decision maker in determine values for the screening criteria during the site screening phase, (b) building the decision model, and (c) assigning weights to the attributes used in the evaluation criteria during the site evaluation phase. A GIS is used to perform the spatial analysis required in the screening phase of candidate sites. The Analytic Hierarchy Process (AHP), a MCDM method, is used to identify the most suitable site in the site evaluation phase. Since the topic of the paper is rather specialized, a brief description of the site selection process, characteristics and tools are described first.

1

Findikaki, "SISES: An Expert System for Site Selection", In M.L. Maher and C.N. Kosten, Proceedings of a Symposium of the American Society of Civil Engineers, Expert Systems in Civil Engineering, 182-192,1986.

2

R.B. Quiambao, "GIS Analysis and Cartographic Presentation of A Site Selection Problem", Proceedings of the 22nd Asian Conference on Remote Sensing, Singapore, 2001.

3

J. Dobson, "A Regional Screening Procedure for Land Use Suitability Analysis", The Geographical Review, 69, 224-234, 1979. doi:10.2307/214965

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £82 +P&P)