Least squares goal programming on the balance set is used to determine the optimal balance point for the corresponding Pareto solution of a multiobjective linear program. Then, standard single objective LP software codes can be used for fast computing of goal optimal Pareto solutions of that multiobjective linear program. The method is applicable for preferential relative cost arrangements (generally non Pareto) and for use with a direct choice of gain/loss ratios for nonimprovable (Pareto) solutions. It will be shown how to use commercially available single objective LP software for computation of goal-optimal or gain/loss preferred Pareto solutions or preferential cost solutions for linear decision problems with multiple objectives. Various illustrative examples will be considered.

All project managers face the problem of coming up with a "good" feasible schedule which take into account project duration, assigned resources and project cost. To help, the project manager could start with specific goals and let a DSS iterate between suggested feasible schedules and adjusted preferred goals. Such an approach has been experimented within an interactive DSS using Tabu search incorporated in a multicriteria approach based on the bounded objective method. Experimentation was performed on a set of published test problems (110 projects). The initial project parameters of original problems were used as preferred goals. With many indicators derived, the experimentation suggests that such a multicriteria approach gives better results than the minimization of a weighted sum of the different objectives for various settings of the weights. The presentation will discuss the approach and present indicators obtained through the experimentation.