Catálogo de publicaciones - libros

In this paper, we discuss our challenge on how to give the creatures and ability to follow spatial restriction while keeping the complexity low enough to still allow for real-time simulation of the herd. Our methodologies extend the pioneering work by Reynolds’ flocking algorithm. We extend how the herd can move in natural-looking paths. Also, we show like the creatures to travel smoothly in 3D space with speed regulation in curve.

In the research of a multiagent system, the indicators such as a task achievement ratio and a payoff have been used for analyzing a system. These indicators are important for the point that agents need to accomplish a task. However they are inadequate to make clear the entity of phenomena that occur in complex system, because they are specialized in the target system and the analysis is also specialized. In these respects the approaches that analyze a system quantitatively are begun to investigate in recent years. In our research, we propose the approach that analyzes a multiagent system quantitatively by focusing on the dynamics of a system and interaction among multiagents. We use two indicators, i.e., entropy and mutual information, for analyzing a system. Entropy estimates the behavior of an agent and mutual information estimates interactions between two agents. For verifying these propositions, we conduct verification experiment in the simple slime mold model. The result shows a relationship between agents’ behavioral patterns and two indicators, therefore our approach using entropy and mutual information is available for analyzing a multiagent system.

On the design issues of power-plant layout engineering, it frequently occurs that objective functions or criteria mutually conflict one another in terms of cost, robustness or delivery time. To this end, we are now developing a concept design support system that is used for plant-layout planning comprehensively by a flexible human-machine communication and multi-objective optimization algorithm. This design approach is based on ‘Shinayaka Systems Principle’, which advocates reactivity, adaptivity and flexibility. In this paper, we discuss the outline of plant-layout planning for power-generating plants, and evaluate suitable multi-objective optimization algorithms including Satisficing Trade-off Method. As a typical example, we apply this method to the model case of combined cycle power plats that are quite popular as an EPC (Engineering, Procurement & Construction) project.

A direct-proportional length-based DNA computing approach for weighted graph problems has been proposed where the cost of each path is encoded by the length of oligonucleotides in a proportional way. During the initial pool generation, the hybridization/ligation phase is carried out where all the combinations are generated in the solution. However, this encoding suffers from biological behavior of hybridization since longer oligonucleotides are more likely to hybridize as oppose to the shorter ones. Recently, parallel overlap assembly (POA) has been recognized as an efficient initial pool generation of DNA computing for weighted graph problems. If POA is employed during the initial pool generation of direct-proportional length-based DNA computing, we expected that the biological influence contributed by various lengths of the oligonucleotides could be minimized as much as possible. Thus, in this paper we found that the hybridization/ligation method should be replaced with parallel overlap assembly, for a better and efficient initial pool generation of direct-proportional length-based DNA computing.

DNA computing approach has gained wide interest in recent years since Adleman shows that the technique can be used to solve the Hamiltonian Path Problem (HPP). Since then there has been many research results showing how DNA computing is used to solve a variety of similar combinatorial problems which is mainly in the realm of computer science. However, the application of DNA computing in solving engineering related problems has not been well established. In this paper we demonstrate how DNA computing can be used to solve a two-elevator scheduling problem for a six-storey building. The research involves finding a suitable technique to represent the DNA sequences in finding the optimal route for each elevator based on initial conditions such as the present position of the elevator, the destination of passengers in the elevators and hall calls from a floor. The approach shows that the DNA computing approach can be well-suited for solving such real-world application in the near future.

A new power system that consists of loosely coupled mini grids is providing a number of challenging problems for many power system experts. In the new development, the power system is said to change from nowadays closely coupled and fully coordinated system that mainly consists of the large power stations, to the mixture of the large power stations and loosely coupled mini grids. The paper focused on the features of the mini grid control system, in particular, its robust generator controller and optimal management schemes.

We regularize Gaussian mixture Bayesian (GMB) classifier in terms of the following two points: 1) class-conditional probability density functions, and 2) complexity as a classifier. For the former, we employ the Bayesian regularization method proposed by Ormoneit and Tresp, which is derived from the maximum a posteriori (MAP) estimation framework. For the latter, we use a discriminative MDL-based model selection method proposed by us. In this paper, we optimize the hyperparameters in 1) and 2) simultaneously with respect to the discriminative MDL criterion, aiming to auto-configure the hyperparameter setting for the best classification performance. We show the effectiveness of the proposed method through some experiments on real datasets.

GA has an advantage that it can treat target functions without depending on their forms. Recently, many studies have been conducted on applying GA into the policy search. Especially, approaches using a rule based policy with Gaussian Mixture are promising. There is not, however, any genetic operator to create a new normal distribution from plural ones. We propose an effective policy representation and a new genetic operator INDX for it. The performance of the proposed method is shown, applying it to two benchmark problems, the Mountain Car and the Cart-Pole Swing up task.

N-version Program (NVP) is a programming approach to fault tolerant software systems. It employs functionally equivalent, yet independently developed software components. We formulate the optimal design problem of NVP system to a biobjective optimization model, i.e., maximizing the system reliability and minimizing the system total cost. We use a Multi-Objective Genetic Algorithm (MOGA) to solve multi-objective optimization problems, however, it requires an appropriate mechanism to search Pareto solutions evenly along the Pareto frontier as many as possible. In our MOGA, we employ the random-key representation and the elitism and Pareto-insertion based on distance between Pareto solutions in the selection process. The proposed MOGA will obtain many Pareto solutions along the Pareto frontier evenly

For estimating the influence of individual’s appreciation experience to the behavior to artwork, we conducted one questionnaire survey and two experiments. In the questionnaires, the depth of concern and knowledge about fine art of each subject were measured. The subjects were resultantly classified to high, middle, and low experience in appreciations of variety of artwork. In the first experiment, eye movements of the subjects while appreciating an abstract artwork were observed. Eyes of experienced viewer moved more widely and captured many characteristic objects of the work, exploring and finding a pattern on the work. In the second experiment, the number of artwork photographs picked by each subject was counted. The high and low experienced subjects picked many pieces from the 46 candidates, whereas the ones of middle experience took time in wondering about their own preference. These results imply the experiences in viewing artwork makes an attention to the meaning of artwork, and that two patterns of preference may occur, i.e., with and without semantic understanding of artwork.