Interactive decision support systems based on linear and non-linear optimization

The paper considers the problem of the multiobjective optimization of a mass transit system. The optimization process concerns the allocation of vehicles to particular routes and defining operating frequencies of these routes. Two mathematical models: I-piecewise linear, non-deterministic and II-non-linear, deterministic of a mass transit are compared. Both models are formulated in terms of multiobjective mathematical programming problems. They contain several criteria which take into account both passengers’ and operator’s point of view. Those models are solved by two different interactive methods, which serve as a basis for construction of two transportation decision support systems. Both methods are characterized by phases of decision alternating phases of computation. At each computation phase the Decison Maker obtains a sample of efficient solutions he/she must analyse and evaluate in the decision phase. The procedures stop when the Decision Maker selects the final compromise solution, satisfying him/her on all objectives.