Fuzzy multiple objective programming techniques in modeling forest planning ∗ Claudia Anderle, Mario Fedrizzi and Silvio Giove Dipt. di Informatica e Studi Aziendali, University of Trento, via Inama 5-7, I-38100 Trento, Italy Robert Fullér † Department of Computer Science, Eötvös Loránd University, Muzeum krt. 6-8, H-1088 Budapest, Hungary Abstract In many cases it is under legislative mandate to manage publicly owned forest resources for multiple uses (i.e., timber production, hunting, grazing). Forest resources have some particular characteristics which make rather difficult their management. In fact they are a typical example of a joint production (market and social goods) and therefore every management policy must face tradeoffs between forest use and forest preservation. In [6] Steuer and Schuler presented a case study of an attempt to apply multiple objective linear programming techniques in management of the Mark Twain National Forest in Missouri. More often than not, accurate market values are not available for some forest products (e.g. dispersed recreation) and, therefore, instead of exact coefficients we have to deal with their approximations (fuzzy numbers) in the modeling phase. In this paper we demonstrate the applicability of fuzzy multiple objective programming techniques for resource allocation problems in forest planning. Keywords: Forest planning, fuzzy multiple objective program, decision support system ∗ in: Proceedings of EUFIT’94 Conference, September 20-23, 1994, Aachen, Germany, Verlag der Augustinus Buchhandlung, Aachen, 1994 1500-1503. † Presently visiting professor at Dipt. di Informatica e Studi Aziendali, University of Trento, Italy. Partially supported by the Hungarian National Scientific Research Fund OTKA under the contracts T 4281, T 14144, 816/1991, I/3-2152 and T 7598 1 1 Introduction Forest planning is a very complex activity because there are many goals which should be achieved simultaneously and a lot of components and elements which must be considered. In fact, it is a typical example of a multicriteria multiperson decision making problem: flexible models must therefore be determined and utilized in order to evaluate the present and future potentialities of the territory and the efficiency of the possible solutions. Furthermore, in forest management one encounters a lot of difficulties e.g., most of the data are not measurable exactly (uncertain or fuzzy); different evaluation criteria and often conflictual expectations; the solution can be unstable under small changes in the imprecise data; tradeoffs between forest preservation and forest use must always be considered (see [2, 4, 5]). It is why one should develop a decision support system, which determines a solution that satisfies these requirements as far as possible. In the forest planning process each decision model distinguishes three main phases: information, analysis and decision. In the information phase the goals, the evaluation indexes (technical and/or logical) of the technically possible alternatives (sylvicolture and/or infrastructure) and the territory potentialities have to be identified. In the analysis phase Pareto-optimal alternatives are searched for by using continuous planning models (Multi Objective Decision Making models) and noncontinuous planning models (Multi Attribute Decision Making models). Due to the peculiarities and limits of both models, it is usually more promising to follow a ”mixed” approach based both on MODM methods (to determine efficient alternatives) and on MADM methods (to determine the most suitable alternative): there are some technical rules being able to ”convert” the effecient alternatives obtained in a mathematical programming model in a multiattribute multiperson decision scheme. In the decision phase we evaluate the decision makers’ preferences and determine an alternative with the highest consensus degree. In this paper we focus our attention on the modeling part of forest management support systems and demonstrate the applicability fuzzy multiple objective programming techniques for the resource allocation problem. 2 2 Formulation of the resource allocation problem The basic problem is to allocate acres and budget monies to alternative management options to meet the best a set of objectives usually specified in goal attainment terms. We can formulate the forestry problem as follows maximize{(C1 x, . . . , C5 x)|Ax ≤ b} (1) where Ci = (Ci1 , . . . , Cin ) is a vector of fuzzy numbers, A is a crisp matrix, b is a vector in Rm and x ∈ Rn is the vector of crisp decision variables. Suppose that for each objective function of (1) we have two reference fuzzy numbers, denoted by mi and Mi , which represent undesired and desired levels for the i-th objective, respectively. Table 1: Objectives, desired and undisered goal levels of attainment Objective Desired goal levels Undisered goal levels Timber production M1 m1 Dispersed recreation M2 m1 Hunting forest species M3 m3 Hunting open land species M4 m4 Grazing M5 m5 We now can state (1) as follows: find an x∗ ∈ Rn such that Ci x∗ is as close as possible to the desired point Mi , and it is as far as possible from the undisered point mi for each i. In multiple objective programs, application functions are established to measure the degree of fulfillment of the decision maker’s requirements (achievement of goals, nearness to an ideal point, satisfaction, etc.) about the objective functions (see e.g. [3, 7]) and are extensively used in the process of finding ”good compromise” solutions. Now we should find an x∗ ∈ Rn such that Ci x∗ is as close as possible to the desired point Mi , and it is as far as possible from the undisered point mi for each i. Let di denote the maximal distance between the α-level sets of mi and Mi , and let mi be the fuzzy number obtained by shifting mi by the value 2di in the direction of Mi . Then we consider mi as the reference level for the biggest acceptable value for the i-th objective function. 3 m1 1 m'1 C1x* C1 x∗ is too far from M1 . It is clear that good compromise solutions should be searched between Mi and and we can introduce weigths measuring the importance of ”closeness” and ”farness”. mi , m1 C1x* M1 C1 x∗ is close to M1 , but not far enough from m1 . Let Ω ∈ [0, 1] be the grade of importance of ”closeness” to the disered level and then 1 − Ω denotes the importance of ”farness” from the undisered level. We can use the following family of application functions [1] Hi (x) = 1 1 + d(Mi (Ω), Ci x) where Mi (Ω) = ΩMi +(1−Ω)mi and d is a metric in the family of fuzzy numbers. 4 M1 ΩM 1 + (1-Ω)m'1 C1x* '1 A good compromise solution Then (1) turns into the following problem max{(H1 (x), . . . H5 (x)) | Ax ≤ b}. (2) And (2) can be transformed into single objective problem by using the minimum operator for the interpretation of the logical ”and” operator max{min(H1 (x), . . . , H5 (x)) | Ax ≤ b} (3) It is clear that the bigger the value of the objective function of problem (3) the closer the fuzzy functions to their desired levels. References [1] C.Carlsson and R.Fullér, Fuzzy reasoning for solving fuzzy multiple objective linear programs, in: R.Trappl ed., Cybernetics and Systems ’94, Proceedings of the Twelfth European Meeting on Cybernetics and Systems Research, World Scientific Publisher, London, 1994, vol.1, 295301. [2] L.S.Davis and G.Liu, Integrated forest planning across multiple ownerships and decision makers, Forest Science, 37(1991) 200-226. [3] M.Delgado,J.L.Verdegay and M.A.Vila, A possibilistic approach for multiobjective programming problems. Efficiency of solutions, in: R.Slowinski and J.Teghem eds., Stochastic versus Fuzzy Approaches to Multiobjective Mathematical Programming under Uncertainty, Kluwer Academic Publisher, Dordrecht, 1990 229-248. 5 [4] P.Kourtz, Artificial intelligence: a new tool for forest management, Canadian Journal of Forest Research, 20(1990) 428-437. [5] M.Kainuma, Y.Nakamori and T.Morita, Integrated decision support system for environmental planning, IEEE Transactions on Systems, Man and Cybernetics 20(1990) 777-790. [6] R.E.Steuer and A.T.Schuler, An interactive multiple-objective linear programming approach to a problem in forest management, Operations Research, 26(1978) 254–269. [7] H.-J.Zimmermann, Fuzzy programming and linear programming with several objective functions, Fuzzy Sets and Systems, 1(1978) 45-55. 6