Operations research methods for forest management and production planning

这项工作涉及在森林管理和生产规划方面应用业务研究方法。森林管理工作涉及一系列综合活动，从长期战略模式到详细的单一林分决策。生产工作主要集中在制造控制上。这些明显不同的领域之间的共同联系主要体现在三个主题上。一种是分层规划方法。在这种方法中，我们强调优化是探索权衡的手段。我们计划开发先进的技术，允许对森林管理中的经济和生态权衡进行分析，类似于我们在生产控制中采用的交换曲线方法。第二个是使用整数规划方法来做出离散决策。整数规划的成功使用需要强有力的公式。我们在邻近地区的森林采伐和鲑鱼水产养殖生产规划方面创造了新的配方。第三个领域是在估计和控制中使用曲面拟合方法。这包括林分水平森林管理中的近似动态规划方法和经济批量方法，以及我们用于制造控制和森林生长建模的神经网络方法。

This work involves the application of operations research methods in forest management and in production planning. The forest management work involves an integrated series of activities ranging from long term strategic models to detailled single stand decision making. The production work is mostly focussed on manufacturing control. The common links between these apparently different areas are in three main themes. One is the hierarchical planning approach. Within this approach we have stressed optimization as means of exploring tradeoffs. We plan to develop advanced techniques that permit analysis of economic and ecological tradeoffs in forest management similar to the use of exchange curve methods that we are pursuing in manufacturing control. A second is the the use of integer programming methods to make discrete decisions. Successful use of integer programming requires strong formulations. We have created novel formulations in the areas of both forest harvesting under adjacency and in salmon aquaculture production planning. A third area is the use of surface fitting methods in estimation and control. This includes approximate dynamic programming approaches in stand level forest management and in economic lot sizing methods as well as the neural network methods that we use for manufacturing control and forest growth modelling.