Abstract

Traditional traffic measurements meter throughput on time scales in the order of 5 minutes, e.g., using the Multi Router Traffic Grapher (MRTG) tool. The time scale on which users and machines perceive Quality of Service (QoS) is, obviously, orders of magnitudes smaller. One of many possible reasons for degradation of the perceived quality, is congestion on links along the path network packets traverse. In order to prevent quality degradation due to congestion, network links have to be dimensioned in such a way that they appropriately cater for traffic bursts on time scales similarly small to the time scale that determines perceived QoS. It is well-known that variability of link load on small time scales (e.g., 10 milliseconds) is larger than on large time scales (e.g., 5 minutes). Few quantitative figures are known, however, about the magnitude of the differences between fine and coarse-grained measurements. The novel aspect of this paper is that it quantifies the differences in measured link load on small and large time scales. The paper describes two case studies. One of the surprising results is that, even for a network with 2000 users, the difference between short-term and long-term average load can be more than 100%. This leads to the conclusion that, in order to prevent congestion, it may not be sufficient to use the 5 minute MRTG maximum and add a small safety margin. Document type: Part of book or chapter of book

Abstract

Traffic management problems provide a unique environment to study how multi-agent systems promote desired system level behavior. In particular, they represent a special class of problems where the individual actions of the agents are neither intrinsically “good” nor “bad” for the system. Instead, it is the combinations of actions among agents that lead to desirable or undesirable outcomes. As a consequence, agents need to learn how to coordinate their actions with those of other agents, rather than learn a particular set of “good” actions. In this chapter, the authors focus on problems where there is no communication among the drivers, which puts the burden of coordination on the principled selection of the agent reward functions. They explore the impact of agent reward functions on two types of traffic problems. In the first problem, the authors study how agents learn the best departure times in a daily commuting environment and how following those departure times alleviates congestion. In the second problem, the authors study how agents learn to select desirable lanes to improve traffic flow and minimize delays for all drivers. In both cases, they focus on having an agent select the most suitable action for each driver using reinforcement learning, and explore the impact of different reward functions on system behavior. Their results show that agent rewards that are both aligned with and sensitive to, the system reward lead to significantly better results than purely local or global agent rewards. They conclude this chapter by discussing how changing the way in which the system performance is measured affects the relative performance of these rewards functions, and how agent rewards derived for one setting (timely arrivals) can be modified to meet a new system setting (maximize throughput). Document type: Part of book or chapter of book

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract