Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.


Share this Page URL
Help

1.5.4 About the Solution Concepts > 1.5.4.9 Comments on the Concept of the Stac... - Pg. 35

1.5 More about the Scope of Game Theory 35 packets, unlike cars, do not choose their routes. These are determined by the network routing protocols. Moreover, for a long time routing issues did not occur in wireless networking, as it was restricted to the access part of a network where each terminal is associated with a given BS. Routing has become relevant to wireless networking since ad-hoc networks were introduced; e.g., see the pioneering work of Gupta and Kumar (1997). However, game theory has penetrated into ad-hoc networks in a completely unexpected way. Indeed, to our knowledge, it was in this context that non-cooperative game theory appeared for the first time as a basic powerful tool for solving problems that do not involve competition. Gupta and Kumar (1997) proposed a routing algo- rithm for packets that had to satisfy some properties: packets should be routed by the network so as to follow the shortest path. It is thus not the average delay in the network that is minimized; the design objectives are to find routing strategies that (i) equalize the delays of packets that have the same source and destination, so as to avoid the re-sequencing delays that are quite harmful for real-time traffic as well as for the TCP protocol for data transfers, and (ii) make these delays minimal, i.e., do not use routes that have delays larger than the minimal ones. The authors designed a network routing protocol that achieved these two objectives. However, these objectives are in fact the WE definition, arising in a completely