Regular Article

Problem-solving using complex networks

¹ Institute of Mathematics and Computer Science, University of São Paulo, São Carlos, SP, Brazil
² Department of Computer Science, Federal University of São Carlos, São Carlos, SP, Brazil
³ São Carlos Institute of Physics, University of São Paulo, São Carlos, SP, Brazil

^a e-mail: h.f.arruda@gmail.com

Received: 22 February 2019
Received in final form: 2 May 2019
Published online: 19 June 2019

Abstract

The present work addresses the issue of using complex networks as artificial intelligence mechanisms. More specifically, we consider the situation in which puzzles, represented as complex networks of varied types, are to be assembled by complex network processing engines of diverse structures. The puzzle pieces are initially distributed on a set of nodes chosen according to different criteria, including degree and eigenvector centrality. The pieces are then repeatedly copied to the neighboring nodes. The provision of buffering of different sizes are also investigated. Several interesting results are identified, including the fact that BA-based assembling engines tend to provide the fastest solutions. It is also found that the distribution of pieces according to the eigenvector centrality almost invariably leads to the best performance. Another result is that using the buffer sizes proportional to the degree of the respective nodes tend to improve the performance.