Transportation Fleet Routing Algorithm Model Based on Parallel Universes (Case Study: Global Distribution of Poultry Meat in Tehran)

Transportation has long been a special place in the economy of Iran and has already had a special position in economic, production and services systems. According to the Central Bank of Iran's statistics, about 13.6 percent of the national income of the country is related to the transport and warehousing sector and about 9 percent of it, belongs to the transport sector. In the past decade, with an average growth of 14.5%, the share of transportation sector has been one of the most important components of economic growth. However, long-standing issues such as the lack of utilization of all fleet capacity and lack of knowledge about optimal distribution routes have caused overhead transportation costs to be off sized by the profits of company’s activities in the distribution sector as well as by manufacturers. Hence, researchers have always sought solutions to improve the transportation routes and eliminate these additional costs.
One of the most important issues in the field of transportation, which is highly considered, is the Vehicle Routing Problem (VRP) issue. The most prominent types of VRP issues is, the carrying goods problem which is limited by the goods delivery time. This issue is very common in carrying perishable goods. The most vital issue in carrying perishable goods such as protein products, is to maintain its quality along the route, which is important in addition to the economic overview in terms of maintaining food safety. Tehran with population around 8293140, is one of the populated city among 25 cities in the world, which has got most important role to study in this issue.
The major part of the protein which is consumed by the citizens of Tehran is poultry meat. In Tehran, about 10% of the physical distribution of poultry meat is made by the grocery section of municipality of Tehran. It is distributed through 141 retail markets, which are supplied by two central warehouses. Accordingly, in the present study, with an examination of the existing transport structure, we will present an optimal pattern for distributing daily chicken meat in Tehran's retail markets.
The VRP refers to a set of issues in which a number of vehicles concentrated in one or more locations should go to a set of customers, each with a specific demand, to provide a service. For the VRP issue, a variety of different constraints are presented. But the limited-capacity vehicle routing (CVRP) is a major example of vehicle routing, in which all customers have the same delivery limited area and specified demand. In the above question, the goal is to minimize the linear composition of the number of paths, the length of the routes or the travel time, so that more convenient and less costly customer service can be provided.
In order to solve the transport problem, two parallel world algorithms and a definite algorithm is used. Then, the algorithms are compared. For this purpose, according to the data of the year 92 of the chicken distribution network in Tehran, a model was made by considering two chicken distribution depots as well as 141 poultry meat market.
The comparison of the answer to the meta-exploratory algorithm is used, and the answer from the definitive methods indicates that the solutions provided by definite methods are superior to the proposed meta-exploratory algorithm and this excellence has come at the expense of using more time. Increasing the execution time of the meta-exploratory algorithm results in the closeness of the solution of the algorithm to the results of definite algorithms. Also, the results showed that the best answer based on definite algorithms was 40231.6 million Rials, which after 36,000 seconds this answer was obtained and during different times, this answer has dramatically improved. Through the algorithm of parallel universes, although the optimal answer is different from the optimal solution of definitive methods (There is only 5.26% difference in results), at the beginning of the problem solution, almost the optimal answer has been obtained, and the longer solving time has not changed much in cost reduction. In addition, by comparing the results of the existing conditions and the optimal transport model, it can be seen that, transportation costs in existing conditions are 2.33 and 2.14 times more than two definitive routing algorithms and parallel universes.
Considering the evidence and research findings, based on the problem-solving time of VRP and the significance of costs, the Tehran municipality can apply the designed model based on parallel algorithm solving methods to improve the poultry meat transportation network. So if all markets in a comprehensive system declare their demand, then distribution of chicken meet can be optimized. It is recommended to create an intelligent system for recording and tracking market orders, for better implementation of the desired transport pattern.