Applied math seminar - Prof. Alfred M. Bruckstein

Speaker
Prof. Alfred M. Bruckstein
Date
19/12/2019 - 15:00 - 14:00Add to Calendar 2019-12-19 14:00:00 2019-12-19 15:00:00 Applied math seminar - Prof. Alfred M. Bruckstein :Title From Ants to A(ge)nts The wonderfully Weird World of Multi-Agent Swarms :Abstract An ant colony is a marvel of cooperation and coordinated, purposeful work carried out by simple a(ge)nts with very limited capabilities. Ants do not have GPS systems, have no compasses nor odometers,  do not use laser range finders , nor do they have good memories or extraordinary computational resources, and employ no sophisticated long-range sensing or communication equipment. Yet they are ruling the earth, by numbers and by resilience, and by some evolution-developed local response algorithms, that rely on pheromone-mediated myopic interactions. The environment becomes a huge, shared resource covered with "chemical memory" signals.   The paradigm of swarm robotics is an attempt to mimic this phenomenal success of nature. In the attempt to analyze the capabilities of colonies of small and limited robots to perform a variety of tasks one encounters formidable mathematical difficulties. The direct problem of analyzing  the emergent global behavior that results from a set of rules of local interaction is tractable in a few interesting cases, like for example in gathering and region covering or patrolling missions, The inverse problem of deriving local rules of behavior, based on the ant-like robots' limited sensing and communication capabilities, is far less approachable. Several examples illustrating the mathematical tools available for analyzing the behavior of swarms of myopic agents will be discussed in my presentation.   Bldg. 216, Room 201 אוניברסיטת בר-אילן - Department of Mathematics mathoffice@math.biu.ac.il Asia/Jerusalem public
Place
Bldg. 216, Room 201
Abstract

:Title

From Ants to A(ge)nts

The wonderfully Weird World of Multi-Agent Swarms

:Abstract

An ant colony is a marvel of cooperation and coordinated, purposeful

work carried out by simple a(ge)nts with very limited capabilities. Ants

do not have GPS systems, have no compasses nor odometers,  do not

use laser range finders , nor do they have good memories or extraordinary

computational resources, and employ no sophisticated long-range sensing or

communication equipment. Yet they are ruling the earth, by numbers and

by resilience, and by some evolution-developed local response algorithms,

that rely on pheromone-mediated myopic interactions. The environment

becomes a huge, shared resource covered with "chemical memory" signals.

 

The paradigm of swarm robotics is an attempt to mimic this phenomenal

success of nature. In the attempt to analyze the capabilities of colonies

of small and limited robots to perform a variety of tasks one encounters

formidable mathematical difficulties. The direct problem of analyzing  the

emergent global behavior that results from a set of rules of local interaction

is tractable in a few interesting cases, like for example in gathering and

region covering or patrolling missions, The inverse problem of deriving

local rules of behavior, based on the ant-like robots' limited sensing and

communication capabilities, is far less approachable. Several examples

illustrating the mathematical tools available for analyzing the behavior

of swarms of myopic agents will be discussed in my presentation.

 

Last Updated Date : 15/12/2019