MULTIOBJECTIVE DESIGN OF DEVELOPMENT PROGRAMS WITH APPLICATION TO THE SAN TIBURCIO REGION, ZACATECAS, MEXICO

Abstract

The formulation of land-use programs fostering the social development of a community is the basic aim of the present work. The San Tiburcio region, located in northern Mexico, provides the background for the design of specific programs as well as a vehicle to present and illustrate the proposed procedure. The advocated program design procedure is made up of four distinct stages; value structuring, project screening, program delineation, and program selection. At each stage a specific technique is proposed in order to achieve the particular output sought. The value structuring phase strives to achieve a characterization of the objective structure, as advocated by a group, and the identification of objective measures useful in the decision making stage. Q-analysis is used to study the structure of the objectives as implied by a given relation among these objectives. The objective structure for the San Tiburcio region is assessed and studied in the context of the proposed value structuring activities. Based upon previous research, a set of land-use projects is formulated by a multidisciplinary team and submitted as potential candidates to integrate a program for the San Tiburcio region. Related to this set of land-use projects is another set of infrastructural and industrial projects proposed by various governmental agencies. A mixed-integer linear multiobjective programming (MILMOP) model is used to handle the determination of specific program profiles. Due to computational and data limitations, a prior screening of alternatives is undertaken, using a knapsack-type model. By considering several types of objectives and levels of resources, a classification of projects is achieved. Two plausible screening criteria are developed to deal with the San Tiburcio problem. The set of screened projects is used as the starting basis of the program delineation stage. A series of program profiles is generated by obtaining various half-compromise solutions of the MILMOP model. Each half-compromise solution is obtained when a particular set of weights is used for the various objective functions. The selection of a given program is undertaken within the context of a multiattribute decision problem. To this end, a technique based on polyhedral dynamics is developed and used to produce a ranking of programs. A series of sensitivity runs are conducted to conclude that only two programs of the five considered are ranked consistently higher than the rest.