A cellular goore game-based algorithm for finding the shortest path in stochastic multi-layer graphs

The shortest path problem in stochastic graphs has been extensively studied, with numerous algorithms proposed using various learning automata models. However, the dynamic nature, diverse individual characteristics, and inherent uncertainties of social interactions necessitate the adoption of stochastic multi-layer social network modeling. This approach provides deeper insights into the complex relationships within social networks. When formulated as a stochastic multi-layer graph, key elements such as the shortest path require redefinition to account for these complexities. This paper explores the shortest path problem in stochastic multi-layer graphs and introduces a novel algorithm based on the Cellular Goore Game (CGG) to address this challenge. The proposed CGG-based algorithm leverages learning automata and extensive edge sampling to determine the optimal path efficiently. By integrating learning automata and selectively sampling from relevant sections of the graph, the algorithm significantly reduces computational complexity. Experimental results on stochastic multi-layer graphs highlight the effectiveness of the proposed algorithm, demonstrating substantial improvements across multiple metrics, including sampling ratio, shortest path ratio, average iterations, and convergence rate.