Yasukuni Mori

Solving the balance problem of massively multiplayer online role-playing games using coevolutionary programming

In massively multiplayer online role-playing games (MMORPGs), each race holds some attributes and skills. Each skill contains several abilities such as physical damage and hit rate. All those attributes and abilities are functions of the characters level, which are called Ability-Increasing Functions (AIFs). A well-balanced MMORPG is characterized by having a set of well-balanced AIFs. In this paper, we propose a coevolutionary design method, including integration with the modified probabilistic incremental program evolution (PIPE) and the cooperative coevolutionary algorithm (CCEA), to solve the balance problem of MMORPGs. Moreover, we construct a simplest turn-based game model and perform a series of experiments based on it. The results indicate that the proposed method is able to obtain a set of well-balanced AIFs more efficiently, compared with the simple genetic algorithm (SGA), the simulated annealing algorithm (SAA) and the hybrid discrete particle swarm optimization (HDPSO) algorithm. The results also show that the performance of PIPE has been significantly improved through the modification works.

Dynamic Analysis of Sea Wave Data Measured by LED Lidar

Form of sea wave is greatly affected by not just flow and depth of sea water, but also wind blowing on the sea surface. Therefore, measurement and analysis of sea wave motion is of assistance for control and operational safety of boats and ships. Generally, oceanic information is gauged by acoustic and electric wave. But these methods have not provided enough spatial and temporal resolution, and are completely out of touch with the on-site needs. Thus, the LED liar for sea wave measurement has been developed. The dynamic analysis of sea wave image measured by the LED lidar was conducted and the relationship with wind speed was evaluated. In this report, we first present the specifications and measurement methodologies of the LED lidar. Then we describe the actual measurements of sea wave with shallow angle by using this lidar and the results of their analysis.

Archive-shared cooperative coevolutionary algorithm using Nash equilibria preservation

Cooperative Coevolutionary Algorithm (CCEA) has been widely applied to solve large scale and complex problems, which result in significant speedups over traditional genetic algorithm. However, CCEA do not search for the optimal solutions but the compromised solutions because of its nature of cooperation. So in the case that the object problem has infinite Nash equilibria, CCEA is expected to converge into suboptimal solution even employing the shared archive. In this paper, we propose a Nash equilibria preservation strategy which enables CCEA to jump out of the suboptimal solution and to keep searching without falling into the same suboptimal solutions found so far. The experiment results show that the capability of the archive-shared CCEA has been improved by using the proposed strategy, especially in dealing with problem having infinite Nash equilibria.

Neural Network Model Generating Symbol Sequence for Songs of Bengalese Finch

Bengalese Finch (shape Lonchura striata var. domestica) sings more sophisticated songs than other birds and the grammar of their songs has been found to be described with probabilistic finite state automaton (PFA). In the present paper, we propose a multilayer neural network model that succeeds in reproducing the qualitatively similar symbol sequence by taking into account the memory process. The use of the present method is illustrated for the songs of Bengalese Finch with particular emphasis on issues of input delay that is necessary to obtain the correct grammar. It is found that the grammar obtained from the simulated symbol sequence using the PFA agrees well with the real one.

Classifier-Independent Visualization of Supervised Data Structures Using a Graph

Supervised data structures in high dimensional feature spaces are displayed as graphs. The structure is analyzed by normal mixture distributions. The nodes of the graph correspond the mean vectors of the mixture distributions, and the location is carried out by Sammons nonlinear mapping. The thickness of the edges expresses the separability between the component distributions, which is determined by Kullback- Leibler divergence. From experimental results, it was confirmed that the proposed method can illustrate in which regions and to what extent it is difficult to classify samples correctly. Such visual information can be utilized for the improvement of the feature sets.