Planering av vägupprustning med hjälp av GIS och optimering : en fallstudie på Holmen Skog

Abstract: The objective of this thesis was to develop a method for the planning of road upgrading at Holmen Skog, a Swedish forest company. A critical issue for creating such plans is to identify roads that are suitable to upgrade to a standard that can handle heavy traffic during spring thaw, since spring thaw is a bottle neck for transporting timber. The method was tested in a case study at a district. The work consisted of two parts. The first was to retrieve and prepare geographical data for the roads. This work was mainly done in a GIS (Geographical Information System) program. The second part consisted of using this data in an optimization problem where the objective was to minimize total upgrading costs for a road upgrading plan. The optimization problem was divided into a hierarchy of two problem levels. For each level, a linear programming problem was formulated. The top level included the entire area, and the lower level was applied for each of a number of subareas. A number of different problems were solved, with different demand constraints, or harvest levels. The result show that it was only profitable to use optimization when harvesting volumes were large, at least 100 000 m3fub. For smaller harvesting volumes, manually created road plans were fully comparable with the optimized road plans. An advantage of using optimization was that the road projects were less scattered within a road system, independently of harvesting volumes. The cost savings of this, considering less moving of construction machinery and harvesting machines, were not included in this work, but if they were it would probably have been shown that the optimization is more profitable also for lower harvest volumes. Furthermore it is obvious that the single most cost reducing factor is road length. The road could therefore be based on road length instead of cost. The benefit of this is that several time consuming steps in the data preparation process could be avoided and also several uncertain calculations. The potential in preparing future road plans are great, both concerning quality and efficiency. Quality could for example be improved by the development of the connection between roads and harvesting areas, in this paper only the forwarding distance has been used. The efficiency could be improved by having road databases ready to use out in the districts.