ABSTRACT:One of the most important decisions a retailer can make is where to locate a retail outlet. Because convenience is so important to today's consumers, a retail store can prosper or fail solely based on its location. Recently, a changing retail environment is augmenting the location importance as retail economic groups develop multi outlet chains of small stores.The methods used in the development and calibration of location models for commercial spaces and sales forecast are multiple, varying from simple analogy forecast models to very complex spatial interactions models, which may incorporate dependence models in a gravitational or logit structure and many exploratory variables. More recent developments incorporating meta-heuristics such as genetic algorithms for the global problem of the multi outlet chain configuration, or the use of Voronoi diagrams in store trade area delimitation, are also presented. Finally, the Geographical Information Systems' role on the decision support process is equally explored.

No abstract

No abstract

No abstract

This paper describes the modelling approach for the sales of gasoline in service stations located in a Portuguese metropolitan area. The models were developed for planning purposes in order to assess potential sales of new sites. The estimates produced are more accurate than those made by previous models and are currently being used to support investment decisions of the largest Portuguese company. Starting from a basic multiplicative regression model, several segmentations of the service stations were obtained. Among these, the submodels for service stations located in urban areas and for service stations located along routes produced a better fit. The models reflect, without exception, the importance of the variables representing the size of service station and the “traffic” passing by, as well as variables which express location and road type. The fast growth experienced by this market explains the lesser impact on sales of the area sales potential and of competition.