Spatial Decision Support Systems for GIS based infrastructure planning

KCS Murty, JL Raheja and Ranjan Pushilal
Central Electronics Engineering Research Institute,
Pilani (Rajasthan)

We are in the process of development of a PC based Spatial Decision Support System (SDSS) with sophisticated analytical capabilities to solve real world problems in public distribution of essential commodities, using resource network analysis and heuristics based solutions under the realm of GIS technologies. The scope of the development is in development of a stand alone software tool for assisting the planners in location of godowns at FCI, Taluka level, Fair Price Shops (FPS) and allocation of these among them. The software design is aimed at working in stand-alone mode under PC based WINDOWS 95 environment. The tool will take into account the dynamic parameters such as supply change, population change, new topologies, seasonal problems as floods, change in local production levels and generate strategies for uniform distribution of items supplemented with better response time, lesser grievances and reduced transportation cost.

The major objectives of the development are:

• To provide Spatial Decision Support System at district level for optimal flow of essential commodities.


• To assist planners in optimal location of new fair price shops, godowns at different hierarchies.


• To assist planners in overall management of commodity distribution under dynamic conditions.

Analytical models for location and allocation do not consider social, economic and political concerns of decision makers. Real world problems have conflicting objectives. The location models fail to search solution space where the utility function has multiple criteria. A decision support system tackles such ill or semi structured problems. It provides solution adaptively by using explicit knowledge of decision makers expert knowledge.

The paper describes an object-oriented approach for the design of a Spatial Decision Support System. The location / allocation models are represented by a set of network objects. Real world objects are represented through inheritance from the network objects. This methodology of system design will help in easy use of the tool for similar other applications.

Page 1 of 1