Complexity and Scenario Robust Service System Design

摘要

The main research idea of this contribution consists in introducing a suitable way of scenario set generation for robust service system design to find the most suitable size of disjoint scenario set. The objective is to achieve acceptably small time demands and sufficient level of robustness respectively. Since the probability distribution of individual scenarios among the whole set is usually not known or it can be hardly estimated, the robustness needs to be evaluated in other way. Here, we suggest making use of the price of robustness to appraise deterioration of the standard objective function as an effect of robustness increase and the coefficient called gain of robustness to asset the robustness improvement. The way of failure scenario set creation follows from previous research focused on revealing those road segments, which are critically important from the viewpoint of system performance. To create a disjoint scenario set, we apply previously developed arc characteristics, by using of which we are able to uniformly distribute given additional transportation performance. By presented research we focus on exploring possible relations among size of scenario set, computational time and robustness characteristics.