A multiobjective dynamic nonlinear robot assignment problem

Abstract

Robots will be used under rapidly changing and highly dangerous circumstances such as rescue operations in a radioactive environment or a fire as well as military operations. The robots are sent to several targets in order to carry out various tasks. The robots we are considering here are able to send and receive messages to and from each other as well as solve nonlinear assignment problems. When the robot salvo is en-route to their targets several events may happen. A number of co- operative robots may get jammed as a consequence of disturbances. Some robots may already have reached their targets. Some robots may not be able to reach all targets. The system being investigated enables the surviving robots to work together in real time and change their pre-set tasks if necessary in order to maximize their effectiveness. In this paper we present a method which solves the reallocation problem using a piecewise linear network algorithm. Experimental results up to 493 targets and 500 robots show that the reallocation of the robots can be done in real time.