Does the Shape of a Territory Influence the Locations of Human Activities? A Numerical Geography Approach

This paper aims at showing how far the shape of a studied area influences the results of optimal location-allocation models. Simulations are performed on rectangular toy-networks with an equal number of vertices but with different length/width ratios. The case of merging two such networks into a common market is also considered. We limit our experience to the Simple Plant Location Problem (SPLP) which captures the fundamental trade-off of economic geography between accessibility and economies-of-scales. Results are analysed in terms of locations, allocations and costs. The results help at understanding how far an area (country/region) has larger development problems than others just because of its shape and/or of the way this area is linked within a common market (elongation of the country and length of the common border). Several real world examples are discussed when interpreting of the results.