Yoshiaki Katada

Estimating the Degree of Neutrality in Fitness Landscapes by the Nei’s Standard Genetic Distance – An Application to Evolutionary Robotics –

In recent years, not only ruggedness but also neutrality has been recognized as an important feature of a fitness landscape for genetic search. As it has been reported that the evolutionary dynamics on a fitness landscape with neutrality is clearly different from the canonical explanations, ruggedness alone might be inadequate describing it. Another measure, i.e., neutrality is required. In this paper, we proposed the use of the Neis standard genetic distance, which originates from population genetics, for estimating the degree of neutrality in fitness landscapes after minor modifications. Several computer simulations were conducted with an evolutionary robotics problem in order to investigate the validity of the proposed approach. The results suggest to us that the Neis genetic distance is a reliable method for estimating the degree of neutrality on real-world problems.

An Approach to Evolutionary Robotics Using a Genetic Algorithm with a Variable Mutation Rate Strategy

Neutral networks, which occur in fitness landscapes containing neighboring points of equal fitness, have attracted much research interest in recent years. In recent papers [20,21], we have shown that, in the case of simple test functions, the mutation rate of a genetic algorithm is an important factor for improving the speed at which a population moves along a neutral network. Our results also suggested that the benefits of the variable mutation rate strategy used by the operon-GA [5] increase as the ruggedness of the landscapes increases. In this work, we conducted a series of computer simulations with an evolutionary robotics problem in order to investigate whether our previous results are applicable to this problem domain. Two types of GA were used. One was the standard GA, where the mutation rate is constant, and the other was the operon-GA, whose effective mutation rate at each locus changes independently according to the history of the genetic search. The evolutionary dynamics we observed were consistent with those observed in our previous experiments, confirming that the variable mutation rate strategy is also beneficial to this problem.

Artificial evolution of pulsed neural networks on the motion pattern classification system

Categorization is one of the most important cognitive abilities for autonomous agents. In natural systems, animals discriminate any object not only by its figure but also by its motion pattern. In this work, we applied the standard GA to evolve pulsed neural controllers for the motion pattern classification system in order to investigate how evolved agents perform the discrimination task, its evolutionary dynamics and the process of self-organization in the neural controllers. The results demonstrate that the agent controlled by the evolved neural networks can discriminate between the objects with the different motion. In the process of evolution, the fitness is improved by the modulation in the connection weights among neurons.

Initial experiments on reinforcement learning control of cooperative manipulations

The paper deals with instance-based reinforcement learning control of autonomous robots. A classifier system, defined in the continuous state and action spaces, is outlined. Based on the sensory state space analysis, we define a learning strategy and fix the structure of the action rules. The classifier system features a nonconservative bucket brigade algorithm and a fast reproduction mechanism. The system developed is then applied to learning cooperative behavior by two robots coupled via a common object, with each robot controlled by its own classifier. The feasibility of this scheme is tested under experiment with two Lynxmotion robots, and a motion pattern of cooperative behavior (lifting up an object) is evolved using the two interacting classifier systems.

Swarm robotic network using Lévy flight in target detection problem

One approach in swarm robotics is homogeneous system which is embedded with sensing, computing, mobile and communication components. In this study, a target detection problem, which is one of navigation problems, was employed. Once a robot detects a target, robots immediately communicate with a base station via intermediate relay robots due to the multi-hop transmission of wireless communication. Therefore, this control task is completed with connectivity of the network. In a target detection problem, we must improve the performance of exploration as well as connectivity of the network. This study investigates the performances of the two types of random walk algorithm in navigation while loosely ensuring connectivity of the robotic network based on our previous study.

Analysis on topologies of fitness landscapes with both neutrality and ruggedness based on neutral networks

Fitness landscapes which include neutrality have been conceptualized as containing neutral networks. Since the introduction of this concept, EC researchers have expected that a population can move along neutral networks without getting trapped on local optima. On the other hand, it has been demonstrated in tunably neutral NK landscapes that neutrality does not affect the ruggedness, although it does reduce the number of local optima. These show that the effects of neutrality are still contentious issues. This paper investigates the effects of neutrality and ruggedness on topologies of fitness andscapes. A neutral network of a fitness landscape is described in a mathematical form based on Harveys original definition with minor modifications. Our results demonstrate that landscapes with a higher degree of neutrality have the larger sizes of neutral networks. For landscapes with the lowest degree of ruggedness, all networks lead to the networks of the highest fitness via any networks. For landscapes with a higher degree of ruggedness, there are few contact points between the networks of high fitness and the ones of the highest fitness, which seem to be isolated, deceptive or rugged.

The Nei's standard genetic distance in artificial evolution

In recent years, not only ruggedness but also neutrality has been recognized as an important feature of a fitness landscape for genetic search. Following that the concept of neutrality in artificial evolution originates from Kimuras neutral theory in natural evolution, it is expected that the dynamics of artificial evolution in the landscapes including neutrality would be described by using techniques in population genetics. Furthermore, new theoretical guidelines might be developed for effective genetic search. In a recent paper, we have discussed the use of the Neis standard genetic distance, which originates from population genetics, for measuring neutrality of fitness landscapes. In our results, several consistencies with the population genetics have been found by applying the Neis standard genetic distance to a tunably neutral NK landscape. Computer simulations are systematically conducted by using a standard genetic algorithm in order to clarify the characteristics of the Neis standard genetic distance. The terraced NK landscape is adopted as a test function.

Tracking the Red Queen effect by estimating features of competitive co-evolutionary fitness landscapes

Open-ended evolution is considered to be caused by several factors, one of which would be co-evolution. Competitive co-evolution can give rise to the “Red Queen effect”, where the fitness landscape of each population is continuously changed by the competing population. Therefore, if such continuous changes are captured, co-evolutionary progress would be measured. In this paper, we estimate features of competitive co-evolutionary fitness landscapes on a predator-prey problem in computer simulations and investigate the Red Queen effect on the fitness landscape. Two types of method were proposed to estimate features, ruggedness and neutrality. One was calculated based on accumulated data so far at each generation, and the other was based on accumulated data during a certain period. The results suggest to us that our method can track the progress of fitness landscapes on competitive co-evolutionary robotics.

Connectivity of swarm robot networks for communication range and the number of robots based on percolation theory

One approach in swarm robotics (SR) is homogeneous system which is embedded with sensing, computing, mobile and communication components. This is identified with mobile wireless sensor networks (WSNs). For some SR tasks, robots need to collect information from the environment and share their data with each other. Due to the multi-hop transmission of WSNs, robots in such networks can communicate with each other via intermediate relay robots. Therefore, it is important to take connectivity of the network into account. This study investigates communication range and the number of robots required for a SR network to achieve connectivity based on percolation theory.

Estimating the Degree of Neutrality and Ruggedness of Fitness Landscapes

In recent years, not only ruggedness but also neutrality has been recognized as an important feature of a fitness landscape for genetic search. As it has been reported that the evolutionary dynamics on a fitness landscape with neutrality is clearly different from the canonical explanations, ruggedness alone might be inadequate describing it. Another measure, i.e., neutrality is required. This study discusses the use of standard genetic distance, which originates from population genetics, for measuring neutrality of fitness landscapes. Firstly, several computer simulations are conducted with a test landscape with neutrality as well as ruggedness in order to clarify the characteristics of standard genetic distance on it. Second, computer simulations are conducted with an evolutionary robotics problem which would be expected to include neutrality in its landscape in order to investigate the validity of the proposed approach on a real-world problem. The results suggest that genetic distance is a reliable method for estimating the degree of neutrality of realworld problems.