Optimal Planning of Wood Harvesting and Timber Supply in Russian Conditions

• Published in: Forests 2020-06, Vol.11 (6), p.662
• Main Authors: Shabaev, Anton, Sokolov, Anton, Urban, Alexander, Pyatin, Dmitry
• Format: Article
• Language: English
• Subjects: block linear programming; decision-support system; multicriteria Boolean nonlinear programming; optimization; timber supply; wood harvesting
• ISSN: 1999-4907; 1999-4907
• DOI: 10.3390/f11060662

This paper describes an approach to the optimal planning of wood harvesting and timber supply for forest companies of Russia. Software and tools successfully used in other countries (e.g., Finland, Sweden, Canada, etc.) are not as effective in Russian conditions for a number of reasons. This calls for the development of an original approach to solve this problem with respect to Russia’s specific conditions. The main factors affecting the operation of wood harvesting companies in Russia were determined. The optimization problem was formulated taking into account all important features of wood harvesting specific to the country. The mathematical model of the problem was formulated and analyzed. An important requirement is that the solution algorithm should find high-quality plans within short computation times. The original problem was reduced to a block linear programming problem of large dimension, for which an effective numerical solution method was proposed. It is based on the multiplicative simplex method with column generation within Dantzig–Wolfe decomposition and uses heuristics to determine feasible solutions based on the branch and bound method. We tested the solution approach on real production data from a forest company in southern Karelia with a planning horizon up to a year. This case study involved 198 sites and 14 machines harvesting up to 200,000 cubic meters from an available stock volume of about 300,000 cubic meters. An increase in profit by 5% to 10% was observed, measured as revenue from the sale of products, net of harvesting and transportation costs.