Fuzzy goal programming in forestry: An application with special solution problems

This paper documents a practical application of fuzzy goal programming in forestry which is extremely difficult to solve using standard linear programming codes on mainframe computers. The application considers selection of the sequence of timber harvests applied over time to a forest. The objective of the fuzzy goal program is to maximize the minimum periodic harvest across all periods. This approach generates solutions with nearly even harvest flows, and with significantly higher present net value than the traditional approach with crisp constraints. The solution difficulty results from the mutually exclusive nature of the harvest decision. The problem arises when many of the fuzzy goals (periodic harvests) are identical and minimal in the current tableau. In this situation, no single pivot step can increase the value of the objective function, and termination generally occurs at a suboptimal solution. Solution performance using several different software products on computers produced by two venders (IBM and UNIVAC (now UNISYS))_are discussed. The problem is solved through numeric tolerance modifications using an algorithm that is relatively less sensitive to termination criteria. © 1991