Tehokkuutta jätehuoltoon : tyhjennysreittien optimointi pääkaupunkiseudulla

Optimization is usually seen as searching and finding the best possible solution or policy for a given problem. Reaching cost effective ways to operate is a key factor to any sector of economy – whether it is private or public sector. The role of economic competitiveness is crucial nowadays. Due to the cost structure of a transportation company, using the best possible routes to distribute goods or such from point A to point B is very important in order to survive in this the competition. Every saved mile will save certain amount of different costs. No matter whether those costs are fixed or varied. The aim is to provide the best service at the least cost. Vehicle routing problem with waste management is based on an old problem such as Travelling salesman problem (TSP) and vehicle routing problem (VRP). There are some special attributes considering optimization of waste collection routes. For example a huge amount of geographically scattered costumer points, fixed capacities of waste hauler vehicles, emptying of vehicles, changing volumes of waste and time windows associated with some costumers. In this research the focus was on optimization of five different waste collection routes using ArcGIS and comparing the results with empirically collected GPS data. The main research questions are following: Is it possible to gain economic and environmental benefits with computerized software route optimization compared to knowledge based routing (now in use)? What are and how great are the possible benefits? What are the factors (real-life) that need to be considered with the route optimization? In order to solve vehicle routing problem, several different algorithms have been developed. An algorithm is an effective method expressed as a finite list of well-defined instructions for calculating a function. Waste collection problems are one of the most difficult operational problems to solve. This is because of the number of possible solutions. Going through all the solutions would take huge amount of computing time. That is why reaching a solution with exact algorithms is usually out of question. Thus heuristics methods are more appropriate. VRP heuristics can be divided into three categories: construction-, improvement- and meta-heuristics. Optimization for the five collection routes were performed with ArcGIS tools. Especially with its network analyst –toolset. ArcGIS network analyst (NA) uses tabu search metaheuristics to find the optimal solution in a network set (Digiroad network). GPS data was collected from the five driven routes and they were compared to ArcGIS NA optimized routes. An excel table was gathered from the results and also the actual transport costs were included for the comparison. ArcGIS NA optimized routes were on average eight percent shorter than the empirical routes. Special importance was on the capacity of the collection vehicle. The greater the capacity the less emptying stops needs to be done. Also scheduling and compactness of collection points have major impact on driven miles and number of emptying of waste bins. Some elements of the real world dynamics attributes are difficult to take into accounts with GIS optimization.