Door-to-Door delivery in Iran - Abstract (1)
On modeling door-to-door parcel delivery services in Iran
Parcel delivery is a complex logistic service, as it serves many small or medium-sized customers who may send or receive parcels. Modeling such delivery system needs to integrate two different research areas of hub location and vehicle routing. As it totally depends on the network and the linkage of the nodes, in this paper, some door-to-door service providers are taken into account to provide suitable information for modeling parcel deliveries of sparse and wide countries. Since the proposed mixed-integer programming model is NP-hard, a new multi-steps solution method based on a simulated annealing algorithm and local search are presented. The results of the proposed model and the solution method are evaluated based on some small test problems. The performance of the solution method is illustrated by solving a real case with all capital cities of 31 provinces in Iran.
A parcel delivery service provides an inexpensive network to transfer parcels between cities (nodes) each of which as a center may send or receive parcels. Since it is not economical to link all nodes of a network to each other, service provider links covered nodes by setting one or more hubs. Hubs are facilities that work as consolidation, connecting, and switching point for flow between local centers (Zanjirani-Farahani et al., 2013). Centers (non-hub nodes) are connected to their nearest hub node through some routes by a truck or larger vehicles. Vehicles collect/distribute all parcels of the non-hub nodes located in their routes and provide links between non-hub nodes and the nearest hub node. As all centers cannot supply a truckload demand, Less-Than-Truckload (LTL) transportation strategy may help to achieve economies of scale in parcel delivery companies (Wasner & Zäpfel, 2004).
To model a parcel delivery system, two different subjects of hub location and vehicle routing problems should be considered. Although the former is a strategic decision and the later is an operational one, researchers believe that these two decisions are strongly linked (Salhi & Rand, 1989) and only the integrated model can provide a reasonable solution for such a complex situation. Furthermore, parcel delivery models have some common features such as: (I) all hubs are connected to each other, (II) nodes are connected to at least one hub, or (III) nodes are connected to the hubs through some tours. However, modeling parcel delivery in each region totally depends on the local features which contravene global standards.
This paper is supposed to examine parcel delivery services and to model the current design of parcel providers of Iran. Although door-to-door parcel delivery has more than 50 years of service record in Iran, it is still an incomplete logistic service and is not able to provide a smooth and steady service for all cities of the country. There are two important reasons for this happening: Iran geographical conditions and imbalance demands of cities. Iran is the 18th-largest countries of the word with the rugged mountain ranges in the west, wide desert basins in the east and long shores in the north and south. So, while the east part contains sparse cities and restricted roads, the other part of the country is more crowded with close cities and complete roads. Besides, not having the proper investment in the east part of the country, in addition to the other reasons, causes imbalance demands between east and other parts of the country.
To prepare enough information for the investigation, some interviews were conducted with managers of two pioneer parcel providers in Iran and their systems were precisely observed for three months. So, the result can be appropriately generalized for other service providers of Iran. It should be mentioned that most passengers’ transportation companies also provide parcel delivery services between cities which have transportation line and some of them with extra charges are even provided door-to-door service. But these companies are not included in this study.
Almost all parcel providers have a hub in Tehran, the most populated and capital city of Iran, while only some providers have more than one hub. Hubs are connected to each other while nodes connect to hubs in stopover routes (as mentioned by Kara & Tansel, 2001); it means that there is no tour between a hub and its allocated nodes. Companies provide delivery services to the cities (branches) in which have some agents. Agents pick up parcels from customer’s places or customer can deliver his/her parcel to the nearest agent place. Each agent delivers its parcels to its city branch in the predetermined time windows. If the parcel amount of a branch is as much as a vehicle capacity, it directly transfers to the nearest hub, otherwise, a vehicle, which may collect the parcels of some near branches, will come to pick up the parcels and transfer them to the hub. All collected parcels will be sorted, consolidated, and allocated to some routine routes. Vehicles leave the hub to distribute the parcels of some distinct branches which located on the route. Delivery and pickup are not simultaneous, so when vehicles reach to the last branch (city), stay for a while (depending on the route between one hour to one day) and then return to the hub while picking up the collected parcels of the visited branches. In fact, each branch is allocated to a hub which is responsible to serve it through a specific route. It means hubs only handle the parcels related to the branches of their routes and transfer all other parcels to the responsible hubs through line haul connections (hub to hub connections).
Delivery time totally depends on the network and which differs from 24 to 72 hours. In all companies, the delivery time between non-hub nodes to the hubs and vice versa is less than 24 hours, as agents call them one-way route parcels. However, the delivery time between two non-hub nodes depends only on the place of nodes, and their routes.
As collecting/distributing parcel inside the cities is done by agents and selecting the routes is totally depends on their experience; it is not the parcel provider concern and only two levels of hub-nodes (routes) and hub-hub (line-haul) connections will be considered in this investigation. Therefore, parcel provider managers are interested in answering the following questions by scientific investigation:
- Where should the hubs be opened to increase the profit of the parcel service?
- Which cities should be considered in the final network and covering which cities is not economical?
- How many routes should be opened for each hub, and what is the best way to allocate branches (cities) to the routes?
- How many vehicles are needed to handle the delivery part of the network system?
The rest of the paper is organized as follows: The next section reviews the related literature. In Section 3 the problem definition and formulation is described in details. Proposed method based on simulated annealing algorithm is explained in Section 4. Experimental results on some small test problems and on a case of Iran road network are presented in Section 5. The last Section discusses the summary and conclusions of the proposed model and solution method.
Ref: Farzad Bahrami, Hossein Safari, Reza Tavakkoli-Moghaddam, Mohammad Modarres Yazdi - Iranian Journal of Management Studies (IJMS) - Vol. 9, No. 4, Autumn 2016- pp. 883-906