Building Optimal Routes for Cargo Delivery in Megacities

Abstract

The main purpose of the study is to develop a method based on the modified ant algorithm and demonstrate the applicability of its use for building of optimal cargo delivery routes, taking into account the real (non-stationary) dynamics of highdensity traffic on the urban road network in megacities. Within the framework of the proposed method the simulation of building optimal routes, by the example of Kiev, was carried out. The results of study indicate the possibility of using the method for the solution to the problems of efficient management of the freight transportation routing process in the conditions of real dynamics of traffic flows (dense traffic, traffic jam, emergency situations, etc.).

The optimization of cargo delivery processes with dynamic route updates in smart logistics

“…Accordingly, the classic ant ACA algorithm and most of its modifications are more universal. This makes it possible to solve routing problems on road networks of cities of required dimensionality [18]. In addition, as shown in [14], ACA has a better optimization effect for DVRP within TSP (lower time and greater solution accuracy) compared to ESA and GA.…”

“…In [18], a modified ant colony algorithm ACAmod was proposed, in which, unlike the ACA and all its known modifications, ant agents along the graph can move asynchronously at certain (even different) speeds. It is also possible to record the results of optimizing a partially traveled path to optimize the further route when changing the edge length of the graph while driving.…”

“…Here, ant agents, as analogs of vehicles, move at appropriate speeds corresponding to the average speeds of TF in these areas [26]. In addition, paper [19] conducted simulation studies of discrete route optimization within TSP using various AI methods: GA, ESA, ACA, and ACА mod [18]. Here, the route optimization was carried out by time and distance on the example of a wide range of test tasks, where the number of nodes varied from 30 to 152.…”

“…Here, the route optimization was carried out by time and distance on the example of a wide range of test tasks, where the number of nodes varied from 30 to 152. It is shown in [19] that ACAmod is the most effective method of optimizing logistics flows in Smart Logistics among the presented AI methods [18]. Its use makes it possible to reduce the time of finding the optimal solution by an average of 15 % and get better results of path optimization in most cases.…”

Emerging Information and Communication Technologies: the Challenges for the Dynamic Freight Management in City Logistics

“…As a rule, vehicles are based on a single depot (warehouse), and the vehicle tours are performed in a single work shift and may include several pickup and delivery points. The optimization process of assigning customers (pickup and delivery points) to trucks and determining the visiting order of customers and routes refer to vehicle routing and scheduling problems (Taniguchi et al, 2001;Ghiani et al, 2003;Russo et al, 2010;Thompson and Zhang, 2018;Danchuk et al, 2019). The basic information needed for the vehicle routing problem (VRP) is the location of customers, road network conditions, travel times, traffic regulations, etc.…”