 |
UCI Networked Systems | |||||||
|
|
| November 6, 2003, 11am, CS 432 "Ad hoc Networking with Swarm Intelligence" Chien-Chung Shen, University of Delaware Abstract: Swarm intelligence refers to complex behaviors that arise from very simple individual behaviors and interactions, which is often observed among social insects such as ants. Although each individual (an ant) has little intelligence and simply follows basic rules using local information obtained from the environment, such as ant's pheromone trail laying and following behavior, (globally) optimized behaviors, such as finding a shortest path, emerge when they work collectively as a group. In this talk, I will present our research on adapting swarm intelligence as a distributed adaptive control mechanism to design multicast routing and topology control protocols for mobile ad hoc networks. Our multicast protocol adapts a core-based approach which establishes multicast connectivity among members through a core node. An initial multicast connection can be rapidly setup by having the core flood the network with an announcement so that nodes on the reverse paths to the core will be requested by group members to serve as forwarding nodes. In addition, each member who is not the core periodically deploys a small packet that behaves like an ant to opportunistically explore different paths to the core. This exploration mechanism enables the protocol to discover new forwarding nodes that yield lower total forwarding costs. I will present simulation results to demonstrate the performance of the proposed approach and to compare it with certain existing multicast protocols. Our topology control protocol employs a distributed approach where each node asynchronously collects local information from neighbor nodes to determine its appropriate transmission power. Its operations do not require any location, angle-of-arrival, topology, or routing information. In particular, the protocol attempts to minimize the maximum power used by any node in the network, or minimize the total power used by all of the nodes in the network. By balancing the positive feedback and the exploration capability of swarm intelligence, the protocol converges quickly to good power assignment with respect to minimization objectives, while adapts well to mobility. In addition, the protocol may achieve common power, or properly assign power to nodes with non-uniform distribution. I will present simulation results and animations to demonstrate the performance of the protocol for different mobility speed, various density, and diverse node distribution. Bio: Chien-Chung Shen received his B.S. and M.S. degrees from National Chiao Tung University, Taiwan, and his Ph.D. degree from UCLA, all in computer science. He is now an assistant professor in the Department of Computer and Information Sciences at the University of Delaware. His research interests include ad hoc and sensor networks, control and management of broadband networks, distributed object and peer-to-peer computing, and simulation. |
| Networked Systems | last modified 5/9/2008 | UCI • NSC • EECS • ICS |