Application of heuristics and fuzzy logic to natural resource modelling

The complexity and dynamics of natural systems poses considerable difficulties for mathematical description. Conventional modelling techniques that rely purely on an analytical, algorithmic approach are poor at capturing nonlinear processes, cumulative effects and uncertainties characteristic of such systems. In this respect, heuristic models using the principles of fuzzy logic are well suited for describing and simulating processes and dynamics of natural systems. Currently, there are few examples of fuzzy knowledge-based models in natural resource management and there is significant potential for future development. This paper demonstrates two successful applications: 1) a fuzzy expert system that applies a novel method of defuzzification to predict changes in the structure, dynamics and mesoscale distribution of fish shoals; and 2) a method utilising fuzzy approximation theory to predict the recruitment of young fish to a fishery based on the parent stock size and past recruitment conditions.