Identifying Regions Based on Flexible User Defined Constraints.

Area of Science:

• Spatial analysis
• Geographic information systems (GIS)
• Computational geography

Background:

• Regionalization is a complex computational and conceptual problem.
• Existing algorithms often require the number of regions to be pre-specified.
• The max-p algorithm offers an alternative by determining the number of regions as an output.

Purpose of the Study:

• To extend the max-p algorithm for increased flexibility in defining region constraints.
• To enhance the algorithm's ability to search the solution space.
• To focus on the multidimensional characteristics of regions.

Main Methods:

• Modification of the existing max-p algorithm.
• Introduction of greater flexibility in defining feasible region constraints.
• Testing with synthetic spatial and attribute data.
• Large-scale computational experiments to determine optimal parameter settings.

Main Results:

• Demonstrated ability to identify regions in maps of varying complexity.
• Identification of rules of thumb for parameter settings.
• Achieved 94% average accuracy in assigning areas to their true regions.
• Nearly 50% of simulations reached 100% accuracy.

Conclusions:

• The enhanced max-p algorithm provides a flexible and accurate method for spatial regionalization.
• The algorithm effectively handles multidimensional region characteristics.
• The identified parameter settings improve solution accuracy across various scenarios.