Title :
Dynamic assignment of multicast groups to publish/subscribe topics
Author :
Holopainen, Visa
Author_Institution :
Dept. of Commun. & Networking, Aalto Univ., Helsinki, Finland
Abstract :
This paper studies the problem of assigning binary vectors to groups in a way that a metric called wastage is minimized. The first 0-1 integer linear programming (ILP) formulation of the problem is given. The problem is studied in the context of assigning multicast groups to publish/ subscribe topics. In this context execution time is a concern. Hence, a fast two-phase heuristic algorithm is presented. Test results show that the first phase of our algorithm finds good assignments much faster than the best existing algorithm (Simulated Annealing). Furthermore, test results give strong evidence that the second phase of our algorithm (assignment change algorithm) converges towards optimum.
Keywords :
integer programming; linear programming; publishing; 0-1 integer linear programming; binary vectors; dynamic assignment; multicast groups; publish-subscribe topics; simulated annealing; two phase heuristic algorithm; Bandwidth; Clustering algorithms; Complexity theory; Context; Heuristic algorithms; Simulated annealing; Temperature measurement;
Conference_Titel :
Computers and Communications (ISCC), 2010 IEEE Symposium on
Conference_Location :
Riccione
Print_ISBN :
978-1-4244-7754-8
DOI :
10.1109/ISCC.2010.5546748