Effects Of The Application Of Staticand Dynamic Approaches Of Pick Up And Delivery Problem In Delivery Processes

The analysis and understanding of big data is leading companies to achieve high levels of strategic planning. The fact that e-commerce is growing so rapidly forces the logistic companies to constantly improve their services. Routing and scheduling problems are key elements of logistic systems. One of the most famous routing problems is Vehicle Routing Problem(VRP), which seeks for an optimal set of routes to deliver or collect some items from the central point called depot to a number of geographically spread cities or customers. There are many variations of VRP, this research concentrates on a particular variant of the VRP known as the Dynamic Pick Up and Delivery Problem with Time Windows(DPDPTW), which is commonly found in the fields of logistics, delivery services and transportation.Past research has mainly concentrated on the static variant of the PDP, where it is possible to schedule all requests prior to the beginning of the service. In case of dynamic problem, the algorithm should react to the arrival of new locations in real time. Time windows are included in the problem when a system is subject to time constraints. This research provides an extensive analysis and discussion on methods to solve static and dynamic PDPs, and study the effects of applying dynamic approaches to static instances and vice versa. Secondary objectives include a complete literature review of PDPTW and DPDPTW. Additionally, this work discusses the logistical issues of delivery services and prospective solution strategies for the problems which lead to more competent delivery systems. An analysis of the state of the art of e-commerce in different parts of the world is also provided.Metaheuristic methods are used to develop the model because of the applicability space. The goal of the simulation is to test the effects of applying dynamic methods to solve static problems which is achieved by, first, validating the dynamic model computationally; second, to set up several problem scenarios and solve them using a dynamic approach and match them with comparable static solutions. The main findings of this thesis support the premise that large static PDP problems can be solved with dynamic approaches with less expensive computational resources.