The article introduces the operational-tactical-strategic context of wood supply in Russian logging companies. In this context, the transition to cut-to-length means that multiple assortments may be produced at multiple locations in the supply chain for numerous destinations. Harvesting-forwarding and transportation are the key activities in this context. This article gives an overview of the mathematical models implemented in a new decision support system (DSS) for these companies. The DSS consists of a number of modules, including optimization of routes for the fleet of harvester-forwarder teams and truck transport. The optimization of routes is based on a two-phase algorithm where a heuristic optimization method is used for choice of best transportation paths, and dynamic programming is used for choice of daily tasks and vehicle routing. The article presents the results of three cases where the DSS has been tested. The cases selected concern daily (operational), monthly (tactical) and yearly (strategic) transport planning where planning horizons were 4 days, 3 months and 1 year, respectively. The studies have all been made in logging companies located in Northwest Russia with annual harvest levels of about 250,000 m3 per company. The studies include from 4 to 9 harvester-forwarder teams, 5 to 24 trucks, 8-10 assortments, 4-129 supply points and 4-5 demand points. The results show that the DSS can be used to support a wide range of planning decisions at company level including truck routing, fleet utilization levels and choice of transport method under new infrastructure assumptions. The potential cost savings calculated range from 14 to 25%.

Key words: cut-to-length harvesting, decision support system, dynamic programming, logistics, Russia