The Political Districting of Kuwait: Heuristic Approaches

• Activity Number: 444
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Government Statistics Section
• Author(s): Shafiqah Alawadhi* and Rym Mahallah
• Companies: Kuwait University and Kuwait University
• Keywords: Constraint propagation ; Integer programming ; Multiple criteria ; Heuristics ; Multiple objective ; Combinatorial optimization

Abstract

This paper models the political districting of Kuwait as a multiple objective combinatorial optimization problem, where each political system is assessed in terms of population and voting equity, geographical contiguity, and social, religious, ethnic, family size, and educational homogeneity. First, it proposes four constructive heuristics and a specialized simulated annealing to generate alternative non-dominated districting plans that may "guarantee" a national consensus. Second, it searches for a districting plan that optimizes a set of criteria (classified as hard and soft constraints) using a tree-search based heuristic. The heuristic takes advantage and combines the orthogonal but complementary strengths of constraint and integer programming. Finally, it compares the proposed solution to both the existing and the previously applied patterns. Thus, this paper offer politicians may a multiple criteria evaluation method that they may apply to choose the most "appropriate" political districting pattern.