Localisation et gestion des livraisons d'une entreprise virtuelle de restauration

这家名为PIDZ的公司位于魁北克地区，已经运营了三年多。它经营餐饮业务，目前拥有三家餐厅。PIDZ的管理人员注意到，他的公司80%的销售额来自送货和外卖订单。他打算发展虚拟餐厅的概念。它包括植入一个或几个厨房，位于战略位置，以覆盖最大的领土，同时提供尽可能短的交付期限。这项研究项目包括实现一个虚拟餐厅并优化其功能。我们的研究主要集中在选址、运输和库存问题上。所讨论的问题是高度组合性的。换句话说，解空间太大了，无法进行详尽的探索。因此，智能搜索的重要性。这项研究的目的是向该公司提出一个有效的软件，允许开设和管理一个由一套厨房和一支送货车队组成的虚拟餐厅。我们亦考虑改善现有食肆的运作，使资源得以有效运用，以期将开支减至最低，并使利润最大化。为了解决这个问题，我们计划使用近似方法。首先，我们将问题转化为一个数学规划问题。将提出几个标准，充其量能捕捉到该公司寻求的不同目标。对于这些标准中的每一个，都将设计和实施一套元启发式算法。还将以软件包的形式提出并实施超启发式算法，以最多指导启发式算法的使用

The company, PIDZ, located in the region of Quebec, has been operational for more than three years. It operates in the field of catering business and possesses presently three restaurants. The administrator of PIDZ noticed that 80 % of the sales of his company result from delivery and take-out orders. He intends to develop the concept of virtual restaurant. It consists in implanting one or several kitchens, located strategically, to cover the biggest territory while offering the shortest possible delivery deadlines. This research project consists in implementing a virtual restaurant and optimizing its functioning. Our study lies in the area of location, transportation and inventory problems. The problem in question is highly combinatorial. In other words, the solution space is too huge to make an exhaustive exploration. Hence the importance of an intelligent search. The objective of this study is to propose to the company an effective software allowing the opening and the management of a virtual restaurant consisting of a set of kitchens and a fleet of delivery vehicles. We also consider improving the functioning of the current restaurants in order to establish an effective use of the resources with the aim of minimizing the expenses and maximizing the profits. To tackle this problem we plan to use the approximation approach. At first, we formulate the problem into a mathematical programming problem. Several criteria, capturing at best different objectives sought by the company, will be proposed. For each of these criteria, a set of meta heuristics will be designed and implemented. A hyper-heuristic algorithm will also be also proposed and implemented, in the form of a software package, to guide at best the use of thesemetaheuristics