Distributed Evolution for Swarm Learning by Dr. Suranga Hettiarachchi
Wednesday, April 22, 10:30 - 11:30P, LF-111
Traditional approaches to designing multi-agents systems are offline, in simulation, and assume the presence of a global observer. In the online, real world, a global observer may be absent, performance feedback may be delayed or perturbed by noise, agents may only interact with their local neighbors, and only a subset of agents may experience any form of performance feedback. Under these constraints, designing multi-agent systems is difficult.
This talk presents a novel approach called "Distributed Agent Evolution with Dynamic Adaptation to Local Unexpected Scenarios" or DAEDALUS to address above issues, by mimicking more closely the actual dynamics of populations of agents moving and interacting as a distributed network of computers in a (task) environment.
Connected Dominating Set in Wireless Ad Hoc Networks by Dr. Grace Yang
Wednesday, March 18, 2 - 3 P, LF-111
Wireless ad hoc networks are infrastructureless multi-hop networks consisting of mobile or stationary wireless devices. They are inherently distributed systems. The popular applications include mobile ad hoc networks (MANETs) and wireless sensor networks (WSNs). The research challenge in this area is the construction of a underlying communication framework. A connected dominating set (CDS) is frequently used in ad hoc networks as a virtual backbone to support efficient routing, service discovery, and area monitoring. With the emergence of new applications and technologies, we need to extend the traditional CDS concept to meet new requirements. This talk introduces the basic concepts in the area of wireless ad hoc networks and also discusses the distributed/localized algorithm design for the communication framework construction in such networks.
“Advancing the Layered Approach to Agent-based Crowd Simulation” by Dr. Ahmed Abukmail
Monday, March 16, 10 - 11 A, LF-111
Crowd behavior simulation has been an active field of research
because of its utility in several applications such as emergency planning and
evacuations, designing and planning pedestrian areas, subway or rail-road
stations, besides in education, training and entertainment. In agent-based crowd
simulations, where each pedestrian is modeled as an autonomous agent, a tradeoff
is commonly made between the complexity of each agent and the size of the crowd.
We adapt a scalable layered intelligence technique from the game industry, for
agent-based crowd simulation. We extend this approach for planned movements,
pursuance of assignable goals, and avoidance of dynamically introduced
obstacles/threats, while keeping the system scalable with the number of agents.
"Achieving self-managed deployment in a distributed environment via utility functions" by Dr. Debzani Deb
Friday, March 13,10 - 11 A, LF-111
This talk presents algorithms and mechanisms for achieving self-managed deployment of computationally intensive scientific and engineering applications in highly dynamic and large-scale distributed environment. The primary focus is on the modeling of the application and underlying architecture into a common abstraction and on the incorporations of autonomic features to those abstractions to achieve self-managed deployment. To accomplish the self-management, a utility-function has been formulated that governs both the initial deployment of an application and maintains the optimality during execution despite the dynamism and uncertainty associated with the application and the networked environment. The deployment decisions are made solely based on locally available information and without costly global communication or synchronization. The self-management is therefore decentralized to provide better adaptability, scalability and robustness.