Effects of Leaf Extracts of Ocimum gratissimum L. on Quality of Fresh Cut Cucumis sativus L.

Abstract

In evolutionary robotics, robot controllers are often evolved in aseparate development phase preceding actual deployment - we callthis off-line evolution. In on-line evolutionary robotics, by contrast,robot controllers adapt through evolution while the robots performtheir proper tasks, not in a separate preliminary phase. In thiscase, individual robots can contain their own self-sufficient evolutionaryalgorithm (the encapsulated approach) where individuals are typicallyevaluated by means of a time sharing scheme: an individual is giventhe run of the robot for some amount of time and fitness correspondsto the robot's task performance in that period.Racing was originally introduced as a model selection procedure thatquickly discards clearly inferior models. We propose and experimentallyvalidate racing as a technique to cut short the evaluation of poorindividuals before the regular evaluation period expires. This allowsan increase of the number of individuals evaluated per time unit,but it also increases the robot's actual performance by virtue ofabandoning controllers that perform inadequately. Our experimentsshow that racing can improve the performance of robots that adapttheir controllers by means of an on-line evolutionary algorithm significantly.