Liu G FEMC2007

Target-Oriented Multiple-Period Simultaneous Forest Planning Using Simulated Annealing Algorithms

Guoliang Liu, Forest Ecosystem Solutions Ltd, 227-998 Harbourside Dr., North Vancouver, BC, V7P 3T2, Canada, gliu@forestecosystem.ca

Integrated forest planning has to consider non-timber resources such as wildlife, biodiversity, water, carbon sequestration and recreation. Professor John Nelson at the University of British Columbia viewed the integrated resource management problem as a puzzle, where various non-timber and timber interests represent the pieces. If these pieces fit together properly, all values are possible to be achieved and maintained; otherwise, some outputs will decrease and a feasible solution may not be achievable. Planning forest is like solving the puzzles. If the forest is planned properly, there is a much better opportunity to maintain all values.

Numerous computer simulation and optimization models have been developed and used for the complicated forest analysis and planning problems. These models have led us to a period of conflict, confusion and argument. Some people believe that simulation is better than optimization because it is fast and generates good results; and some other people think that optimization is better than simulation because it has intelligence and produce better results.
Both simulation and optimization have their own advantages. I will introduce the Forest Simulation Optimization (FSOS) model that uses both simulation and optimization algorithms Simulated Annealing to arrange and schedule forest treatments such as reforestation, spacing, fertilization, thinning, pruning, partial cutting and final harvest. The simulation schedules one period at a time based on constraints. Simulated Annealing optimization schedules all periods simultaneously to transform forests to their desired states while attempting to maintain the states, maximize, sustain and balance multiple resources for all periods of the whole planning horizon. The desired states are defined by multiple sustainable ecosystem objectives that include water quality, wildlife habitat, biodiversity, carbon sequestration, visual quality, recreation, timber production and economics.

In this presentation, I will demonstrate the differences between the constraint-based period-by-period time-step simulation and the targeted-oriented multiple-period simultaneous optimizations used in the Forest Simulation and Optimization System (FSOS) model.