Dynamic and stochastic vehicle routing problem: a simulation tool for a large-scale study case.

Logistic management is an increasingly topical issue nowadays. Most enterprise try to improve their economics by optimizing their supply chain, continuously reducing costs in the face of improving performance. The current market is mostly made up of small and medium-sized enterprises (SMEs) not used to thinking in term of global optimum and often do not even have the internal skills necessary to make their system efficient. Over the past twenty years globalization has destroyed the boundaries of local markets, as well the single European market brought to overcome the traditional vision of “local-market”. Internet has totally changed consumer habits. The e-commerce, born as an opportunity, is now a consolidated reality. Some large companies have begun to look abroad with a view to growth, others in order to survive in saturated national markets. All this has resulted in several changes in these company: new strategies, need for new management, birth of new roles. Logistics have become the key to success in industries, supply chain management a key function in order to minimize cost or maximize profits. In this work we will discuss a technical problem related to a famous logistic theme, the vehicle routing problem (VRP) in one practical exception. The purpose is to demonstrate which and how many benefits are brought by an engineering approach to problems. Let’s imagine a small furniture shop with few employees and not-owned furniture factory. The shop decides to sell online using a proprietary website. The demand is low, and shipments are managed by the shop itself. For this type of products is needed an attended delivery, in other worlds the customer has to be at home the day of the delivery. This small shop promise to customers a due date by which they will receive the goods bought. Moreover, the furniture shop has not an ERP system to manage neither employees nor production. This situation is common to small reality due to cost and human skills lack. Since the shop decide to sell from their e-commerce platform now receives orders from a much larger area than before. This growing market brought a totally new level of complexity for the daily operations of the shop, historically shipments were not a problem and were organized manually but with a growing demand it became impossible. A simulation tool will be developed in order to represent in a realistic way this scenario, and then solve it. Moreover, to address the specific problem will be developed a likelihood function capable of optimizing deliveries.