Symbiosis in the natural world is of great evolutionary significance. The concept of symbiosis has been applied for designing evolutionary algorithms. In this paper, symbiosis is considered an operator instead of a state, and symbiotic association is regarded as a result of the emergent cost-benefit properties during an evolutionary process. A coevolutionary individual-based metapopulation model of an endosymbiont-host system is introduced in which two species, the potential host and parasite, are assumed. All the subpopulations in the model are represented by NKC models. A random-parasitism model is applied for representing symbiotic associations between a parasite and a host with no predefined intrinsic cost-benefit property. The symbiostic associations between hosts and parasites are considered genetic relationship through which genome of the parasite is transferred to the host. We investigated the coevolutionary dynamics through computer simulations and found the qualitative conditions where symbiosis stably emerges in the model.
This paper studies the robustness analysis of feedback systems for two-wheeled vehicle systems which are designed based on generalized canonical transformations. Firstly the relation between the design parameter in the stabilization procedure and the sensitivity against external disturbances is discussed. Secondly a sufficient condition for the robust stability against unknown physical parameter variations is derived. Finally the effectiveness of our method is demonstrated by experiments.
Many machine learning methods have been proposed to learn techniques of specialists. A machine has to learn techniques by trial and error when there are no training examples. A reinforcement learning is a powerful machine learning system, which is able to learn without giving training examples to a learning unit. But it is impossible for the reinforcement learning to support large environments because the number of if-then rules defined by combinations of a relationship between one environment and one action becomes huge. In a previous paper, we proposed a new reinforcement learning with fuzzy evaluation environment, called FEERL (Fuzzy Environment Evaluation Reinforcement Learning). The FEERL is made up from a fuzzy evaluation, an environment simulator and a search. It was applied to the chess and its effectiveness was confirmed. In this paper, we apply the FEERL to LightsOut game having no opponent as an example of huge environment and show that the FEERL avoids detour actions in search and then get a proper solution.
In this paper, the issue of constructing piecewise linear Lyapunov functions (PWLLF) for stability analysis of nonlinear systems satisfying generalized sector conditions is addressed. The existence condition of PWLLF is represented by a linear programming problem (LP in brief). If the optimal value of the LP is negative, then PWLLF is constructed by using the optimal solution of it. When the optimal value of the LP is nonnegative, the candidate of PWLLF is modified with additional freedom and a new LP is formulated corresponding to the new PWLLF candidiate. It is shown that the optimal value of the resulting new LP is always less than or equal to that of the old LP. The main purpose of this paper is to propose a fast method to solve LPs which appear repeatedly in constructing PWLLF. The method is based on a kind of sensitivity analysis approach of LP and utilizes the special structure of LPs that appears in constructing PWLLF.
Evolutionary Programming (EP) belongs to a class of general optimization algorithms based on the model of natural evolution. EP has also been applied to real-valued function optimization since the early 90's. However, recent research results have proved that EP is not so robust as expected; EP performs very well only when the lower bound of strategy parameters is adjusted to each problem. In order to overcome this difficulty, an extended EP, called Robust EP (REP), is proposed. A major feature of REP is that genetic drift is introduced as another source of changing strategy parameters. Computer simulations are conducted in order to illustrate the robust performance of REP against the lower bound on a set of popular benchmark problems. Some evolutionary characteristics of REP are also clarified by calculating basic statistical values.