Presentation Information
[B-14-20]A Multi-Criteria Weighting Model for Supporting Trade-off Decision-Making
◎Yuka SUGIMOTO1, Sayaka Yagi1, Katsuya Minami1 (1. NTT)
Keywords:
Multi-Criteria Decision Making,Weighting,Autonomous Network Management
Toward autonomous network operations, it is important to select an appropriate action from multiple candidates according to the situation. For example, when traffic increases, possible actions include waiting, traffic rate control, and shifting traffic to a secondary route. In this case, maintaining QoE, suppressing congestion on the current route, and reducing impacts on other routes become decision criteria. Since the importance of these criteria changes with the situation, it is difficult to flexibly determine whether and when to execute an action using only predefined rules or thresholds. Such action selection can be regarded as trade-off decision-making in network operations.This paper investigates a multi-criteria weighting model to support such decision-making. The model uses situational information, such as event type, impact scope, future deterioration, available time, and information reliability, as input. Based on this information, it derives situation-specific decision criteria, assigns weights, and compares candidate actions. This enables the system to present the recommended action and each criterion’s contribution to the result. By updating weights with situational changes, the model is expected to support operational decisions that are difficult to handle with fixed rules.