Abstract (EN)

This paper develops a method for interactive MultiObjective NonLinear Programming procedure (MONLP). It provides a detailed description of an efficient algorithm, and reports on promising computational results. It also discusses several alternative strategies for implementing GRG code (Generalized Reduced Gradient), which is known as one of the ‘best’ methods for solving NonLinear optimization Problems (Abadie, 1978). The method relies on three steps: 1) generation of a subset of feasible efficient solutions; 2) interactive definition by Decision Maker (DM) of his preference structure according to desired outcome; 3) determination of a compromise solution using nonlinear optimization; a global analysis based on “reference point search procedure” is performed (in the criteria space). Following this methodology, it is possible for the DM to find his final solution. A microcomputer version (for medium problems) of the method is available.