Application of Ant Colony Optimizer (ACO) For Effective Path Planning in a Big-Box Store or Retail Facility

ABSTRACT
Real-life stochastic problems are better addressed by adopting metaheuristic techniques. One of the interesting metaheuristic techniques for defining the shortest path is the ant colony optimization (ACO) algorithm. A considerable number of maps for shortest path have been considered in time past using classical techniques which is appropriate for deterministic variables. For stochastic or nondeterministic decision variables, metaheuristic techniques are much more appropriate. This is possible by mimicking the path navigation and swarm propensities of natural entities to provide real-time quality geographical images representing diverse areas or terrain for easy access to routing and path planning for sustainability and economic benefits in systems. In this research, the solution power of ACO has been demonstrated to predict customers’ behaviour in a popular retail outlet, using the travelling salesman problem (TSP) for stochastic shortest path during the purchase of items in a big-box facility with multiple products and sixteen (16) sections. Data obtained from the facility has been validated. The tour length was subjected to pheromone optimization to obtain a pheromone update of 0.00345 per metre as the maximum and 0.001725 as the best update at varying evaporation rate. In conclusion, out of the selected sections, two major paths in the big-box facility yielded optimal tour length and as such either of the paths can be followed by customers to spend the minimum required time in the
facility