Jointly optimizing both the body and brain of a robot is known to be a challenging task, especially when attempting to evolve designs in simulation that will subsequently be built in the real world. To address this, i...
详细信息
Jointly optimizing both the body and brain of a robot is known to be a challenging task, especially when attempting to evolve designs in simulation that will subsequently be built in the real world. To address this, it is increasingly common to combine evolution with a learning algorithm that can either improve the inherited controllers of new offspring to fine tune them to the new body design or learn them from scratch. In this article an approach is proposed in which a robot is specified indirectly by two compositional pattern producing networks (CPPNs) encoded in a single genome, one which encodes the brain and the other the body. The body part of the genome is evolved using an evolutionary algorithm (EA), with an individual learning algorithm (also an EA) applied to the inherited controller to improve it. The goal of this article is to determine how to utilize the results of learning process most effectively to improve task performance of the robot. Specifically, three variants are investigated: 1) evolution of the body+controller only;2) a learning algorithm is applied to the inherited controller with the learned fitness assigned to the genome;and 3) learning is applied and the genome is updated with the learned controller, as well as being assigned the learned fitness. Experiments are performed in three different scenarios chosen to favor different bodies and locomotion patterns. It is shown that better performance can be obtained using learning but only if the learned controller is inherited by the offspring.
Memetic computation is a paradigm that uses the notion of meme(s) as units of information encoded in computational representations for the purpose of problem-solving. It covers a plethora of potentially rich meme-insp...
详细信息
Memetic computation is a paradigm that uses the notion of meme(s) as units of information encoded in computational representations for the purpose of problem-solving. It covers a plethora of potentially rich meme-inspired computing methodologies, frameworks and operational algorithms including simple hybrids, adaptive hybrids and memetic automaton. In this paper, a comprehensive multi-facet survey of recent research in memetic computation is presented.
evolutionary Robotics is a powerful method to generate efficient controllers with minimal human intervention, but its applicability to real-world problems remains a challenge because the method takes long time and it ...
详细信息
evolutionary Robotics is a powerful method to generate efficient controllers with minimal human intervention, but its applicability to real-world problems remains a challenge because the method takes long time and it requires software simulations that do not necessarily transfer smoothly to physical robots. In this paper we describe a method that overcomes these limitations by evolving robots for the ability to adapt on-line in few seconds. Experiments show that this method require less generations and smaller populations to evolve, that evolved robots adapt in a few seconds to unpredictable change-including transfers from simulations to physical robots- and display non-trivial behaviors. Robots evolved with this method can be dispatched to other planets and to our homes where they will autonomously and quickly adapt to the specific properties of their environments if and when necessary.
The interaction between evolution and learning called the Baldwin effect is a two-step evolutionary scenario caused by the balances between benefit and cost of learning in general. However, little is known about the d...
详细信息
The interaction between evolution and learning called the Baldwin effect is a two-step evolutionary scenario caused by the balances between benefit and cost of learning in general. However, little is known about the dynamic evolution of these balances in complex environments. Our purpose is to give a new insight into the benefit and cost of learning by focusing On the quantitative evolution Of phenotypic plasticity under the assumption of epistatic interactions. For this purpose, we have constructed an evolutionary model of quantitative traits by using an extended version of Kauffman's NK fitness landscape. Phenotypic plasticity is introduced into Our model;whether each phenotype is plastic or not is genetically defined, and plastic phenotypes Can be adjusted by learning. The simulation results clearly show that drastic changes in roles of learning cause three-step evolution through the Baldwin effect and also cause the evolution Of genetic robustness against mutations. We also conceptualize four different roles of learning by using 2 hill-Climbing image of a population On a fitness landscape.
In this article we describe a methodology for evolving neurocontrollers of autonomous mobile robots without human intervention. The presentation, which spans from technological and methodological issues to several exp...
详细信息
In this article we describe a methodology for evolving neurocontrollers of autonomous mobile robots without human intervention. The presentation, which spans from technological and methodological issues to several experimental results on evolution of physical mobile robots, covers both previous and recent work in the attempt to provide a unified picture within which the reader can compare the effects of systematic variations on the experimental settings. After describing some key principles for building mobile robots and tools suitable for experiments in adaptive robotics, we give an overview of different approaches to evolutionary robotics and present our methodology. We start reviewing two basic experiments showing that different environments can shape very different behaviours and neural mechanisms under very similar selection criteria. We then address the issue of incremental evolution in two different experiments from the perspective of changing environments and robot morphologies. Finally, we investigate the possibility of evolving plastic neurocontrollers and analyse an evolved neurocontroller that relies on fast and continuously changing synapses characterized by dynamic stability. We conclude by reviewing the implications of this methodology for engineering, biology, cognitive science and artificial life, and point at future directions of research. (C) 1998 Elsevier Science Ltd. All rights reserved.
We propose a modular architecture for autonomous robots which allows for the implementation of basic behavioral modules by both programming and training, and accommodates for an evolutionary development of the interco...
详细信息
We propose a modular architecture for autonomous robots which allows for the implementation of basic behavioral modules by both programming and training, and accommodates for an evolutionary development of the interconnections among modules. This architecture can implement highly complex controllers and allows for incremental shaping of the robot behavior. Our proposal is exemplified and evaluated experimentally through a number of mobile robotic tasks involving exploration, battery recharging and object manipulation.
Existing Recurrent Neural Networks (RNNs) are limited in their ability to model dynamical systems with nonlinearities and hidden internal states. Here we use our general framework for sequence learning, evolution of r...
详细信息
ISBN:
(纸本)1595930108
Existing Recurrent Neural Networks (RNNs) are limited in their ability to model dynamical systems with nonlinearities and hidden internal states. Here we use our general framework for sequence learning, evolution of recurrent systems with LINear Outputs (Evolino), to discover good RNN hidden node weights through evolution, while using linear regression to compute an optimal linear mapping from hidden state to output. Using the Long Short-Term Memory RNN Architecture, Evolino outperforms previous state-of-the-art methods on several tasks: 1) context-sensitive languages, 2) multiple superimposed sine waves.
暂无评论