Journal of Advances in Mathematics and Computer Science

Across the world, and particularly in Africa, elections are increasingly marked by protest movements, often resulting from a sense of injustice or from the contestation of results. It therefore becomes essential to adopt a voting system that promotes collective acceptance and reflects a general consensus. Indeed, it is not uncommon for certain candidates to refuse to acknowledge their defeat, thereby fueling post-electoral tensions. In this context, the VDEPCI method (Vote based on a Distance for Evaluating Preferences with respect to the Ideal Candidate) has been proposed as a promising alternative. It aims to minimize divergences by taking into account both voters’ preferences and a confidence index in the rankings. However, applications of this method have so far been limited to simple cases, involving a small number of candidates and voters. Our research work is part of a perspective of extending this method to more complex electoral contexts, with a larger number of candidates and voters (several thousands), by emphasizing algorithmic implementation and the optimization of the computations required for its application. The main computational result establishes that the proposed implementation has a linear time complexity, O(n×m), with respect to the number of voters and candidates. Experimental
results confirm excellent large-scale performance, demonstrating that the method can efficiently handle elections involving several thousands of voters while maintaining very low execution times.