Modeling Urban Land-Use Suitability with Soft Computing: The GIS-LSP Method

Multicriteria Evaluation (MCE) is a commonly used approach for creating suitability maps in a raster-based Geographic Information System (GIS) computing environment. MCE aggregation can be performed by a linear additive model known as the Weighted Linear Combination (WLC) or Simple Additive Scoring. The WLC uses normalized weights that allow decision makers to model the importance of various criteria relative to each another. However, all WLC methods exhibit significant limitations. As the number of inputs increase, the significance of each input decreases leading to a data loss condition. WLC assumes equal and neutral logical conditions for all inputs, yet in many cases inputs must model conjunctive, disjunctive, mandatory, sufficient, or optional requirements. These problems compromise the WLC level of detail and justification required for structuring highly complex suitability problems such as land-use suitability. The Logic Scoring of Preference (LSP) approach is presented as an alternative to conventional WLC models. The LSP is a soft computing approach comprised of methods for structuring decision problems. The approach features an efficient nonlinear method for suitability aggregation and allows the expression of a spectrum of logic requirements. The goal of this study is to integrate the LSP method into a raster-based geographic information system, with the specific objectives to test the developed approach within a GIS framework using realistic geospatial datasets. The GIS-LSP approach is applied to an urban land suitability problem using raster GIS data for the Bowen Island Municipality, Canada. The results indicate that the integrated LSP-GIS approach is a flexible and robust tool for land-use suitability analysis. The capability of the LSP to perform aggregation and model logical requirements in the decision making process indicate that the LSP-GIS outperforms the WLC-GIS analysis with respect to selectivity and level of detail.