Ant Colony Optimization Algorithm as a Means of Efficiency Improvement for the Last Mile Delivery in Internet Retailing

Rate this item
(0 votes)

MAGAZINE №4(87) August 2018

AUTHORS 

VINOGRADOV A.B.

YUNEEVA D.R. - School of Logistics, National Research University Higher School of Economics (Moscow, Russia)

CATEGORY E-commerce corporate logistics Optimization and mathematical modelling Transportation in logistics

ABSTRACT

Growth rates of internet retailing in Russia outperform brick-and-mortar segment, which raises attractiveness of e-commerce for new players. However a growing number of newcomers make e-tailers seek new competitive advantages and pay specific attention to the logistics support of their businesses. Last mile delivery tends to be one of the most important, though also problematic logistics processes in online retailing.  Potential area of improvement for this process involves application of heuristic routing methods.  These methods allow to find a nearly optimal solution with substantially lower cost of resources compared to traditional methods.

The paper focuses on the heuristic method of a travelling salesman problem solution complicated by the specifics of internet retailing (big number of clients and, hence, delivery points). This method is based on the simulation of ants’ behaviour seeking the shortest path between their colony and the source of food. The authors describe a mathematical model of an ant colony optimization algorithm (ACO) and review its basic steps using the numerical example. Steps of the ACO include definition of the number of nodes, distance between them as well as pheromone concentration; location of couriers (delivery vans) in the nodes; identification of the probability of moving from the initial point (node) to all other points; selection of the movement direction; repetition of the preceding steps (apart from the initial one) for a new node and for the following ones up to the end of the cycle; pheromone renewal; accomplishment of the next cycles (iterations); finding the shortest delivery route. Comparative analysis has shown major ACO benefits including fast solution of high-dimensional problems and algorithm applicability for the non-stationary systems with changing parameters (much resembling online retailing). An opportunity to apply ACO for the last mile delivery routing referring to the vast majority of e-tailers will significantly depend on the speed of development and proliferation of the respective software as well as on improving of selection and adaptation of the algorithm fine-tuning parameters.         

 Electronic version

 Keywords:  

Read 360 times

Contacts

Postal address:

125 319  Chernуakhovskogo str.16

phone/fax (495) 771 32 58

Working with authors: Levina Tamara Vladimirovna

mob. 8-962-965-48-54

E-mail: levina-tamara@mail.ru

ISSN 2587-6775

Published since 2004

Journal's Indexing