The goal of this research is to understand the true distribution of character patterns. Advances in computer technology for mass storage and digital processing have paved way to process a massive dataset for various pattern recognition problems. If we can represent and analyze the distribution of a large-scale pattern set directly and understand its relationships deeply, it should be helpful for improving classifier for pattern recognition. For this purpose, we use a visualization method to represent the distribution of patterns using a relative neighborhood graph (RNG), where each node corresponds to a single pattern. Specifically, we visualize the pattern distribution using a compressed representation of RNG (Clustered-RNG). Clustered-RNG can visualize inter-class relationships (e.g. neighboring relationships and overlaps of pattern distribution among “multiple classes”) and it represents the distribution of the patterns without any assumption, approximation or loss. Through large-scale printed and handwritten digit pattern experiments, we show the properties and validity of the visualization using Clustered-RNG.
: Renewable-Energy Distributed Generation units (REDGs) that are installed in distribution network offer a lot of advantages in terms of technical, economic and environmental benefits. However, REDGs have both positive and negative effects depending on their sizes and locations. Thus, the purpose of this study is to determine the optimal locations and sizes of REDGs that attains fuel consumption reduction and system reliability improvement while satisfying various constraints and considering relevant uncertainties. To optimize the locations and sizes of REDGs, Adaptive Clonal Differential Evolution (ACDE) is proposed to improve the performance of Clonal Differential Evolution (CDE) by updating the control parameters in an adaptive manner. Previously, CDE with randomized scaling factor was introduced. CDE algorithm is capable of enhancing the exploration and searching ability, hence accelerates the convergence of the algorithm. However, the randomized scaling factor does not guarantee the robustness of the algorithm. Therefore, control parameter adaptation that utilizes collected data is introduced to favour providing information on good parameter values. The proposed algorithm is verified on a 33-bus test system. The comparative studies are carried out and the simulation results show that the proposed algorithm is more stable and robust than CDE. a load flow analysis for each of for geographical
Reinforcement learning (RL) enables an agent to find an optimal solution to a problem by interacting with the environment. In the previous research, Q-learning, one of the popular learning methods in RL, is used to generate a policy. From it, abstract policy is extracted by LVQ algorithm. In this paper, the aim is to train the agent to learn an optimal policy from scratch as well as to generate the abstract policy in a single operation by LVQ algorithm. When applying LVQ algorithm in a RL framework, due to an erroneous teaching signal in LVQ algorithm, the learning sometimes end up with failure or with non-optimal solution. Here, a new LVQ algorithm is proposed to overcome this problem. The new LVQ algorithm introduce, first, a regular reward that is obtained by the agent autonomously based on its behavior and second, a function that convert a regular reward to a new reward so that the learning system does not suffer from an undesirable effect by a small reward. Through these modifications, the agent is expected to find the optimal solution more efficiently.
A new stability criterion for bidirectional DC-DC converters in a dc power system is presented in this paper. This proposed criterion gives an overall indicator for the stability at each node in the dc power system. It based on perturbing the system at a certain node by injecting a small ac current at this node, and accordingly monitoring the variation of the voltage at this node. Then, the node impedance can be calculated. This node impedance implies the stability status of the system at that node. The concept of the node impedance stability criterion is introduced, and the application of this criterion to a dc power system is investigated, as well.
Explosive spread of wireless local area network (WLAN) access points in a global range makes this class of wireless system to be one of the best alternative ways to access internet. This paper investigates how many WLAN access points (APs) are distributed in real world and estimates system capacity if they are operated as a single system. The results indicate that the WLAN has a potential to achieve significant high system capacity compared to the existing cellular systems.
Integration of Ge on the Si platforms is essential for the development of next generation large-scale integrated circuits. The Ge-on-insulator (GOI) structure is suited for the realization of high-mobility transistors channels and as epitaxial templates for optoelectronic and spintronic materials. In this work, the fabrication of thin (~50 nm) (100) GOI by the rapid melting growth process has been investigated. Growth with unstable crystal orientation has been observed in wide (=1 µm) GOI strips. However, orientation stabilized growth was achieved in narrow strips (~0.5 µm). Further stabilization of growth orientation was observed in mesh patterned growth with GOI width of 1 µm. Epitaxial growth of Ge was performed on the above structures and the formation of uniform epitaxial layer was demonstrated.
Scanning Hall-probe microscopy is a very powerful technique for the characterization of local properties in superconducting tape. On the other hand, the quantitative results can only be obtained with a known distance between the Hall-sensor and the sample. In this study, we have proposed an estimation method of such a distance without any gap sensor. Considering the boundary condition that current should be zero outside of the sample, we have succeeded in estimating such a distance only from a measured distribution of magnetic field. This method will be a key technology for the characterization of long superconducting tapes by reel-to-reel magnetic microscopy where there would be some possibility of fluctuation of the distance due to vibration at a high-speed measurement.
(111)-oriented Ge-on-insulator (GOI) is the key material structure for next generation multifunctional large scale integrated circuits. The (111) GOI structure can be implemented for high-speed transistor channels, as well as templates for the integration of optoelectronic and spintronic materials on the Si platform. The rapid melting growth technique is an effective way to obtain high-quality GOI structures on Si substrates. However, in formation of GOI strips (width: ~3 µm, thickness: 100 nm) from Si(111) seed, rotation of crystal orientation occurs along growth direction. In this study, we investigate the effects of GOI pattern-dimensions on orientation stability and demonstrate the suppression of crystal rotation by narrowing the strip width. This enables the formation of (111) GOI strips with any growth direction.
Totalizer (TO) by Bailleux et al. and Half Sorting Network (HS) by Asin et al. are typical CNF encoding methods of cardinality constraint. The former is based on unary adder, while the latter is based on odd-even merge. Although TO is inferior to HS in terms of the number of clauses, TO is superior to HS in terms of the number of variables. We propose a new method called Modulo Totalizer (MTO) to overcome the disadvantage of TO. As an application, we have developed a partial MaxSAT solver with MTO. Preliminary experimental results show that our MTO based MaxSAT solver is comparable to or surpass the conventional TO based maxsat solvers.
A practical method is proposed for forecasting global solar radiation in a wide area which, in turn, is used for photovoltaic generation output forecast. For ease of forecasting procedure, the area is divided into sub-areas or groups of sites. Sites having similar weather form a group. The similarity is calculated using mean absolute deviations or correlations among the actual sunshine duration data. The forecasting procedure becomes simpler by focusing on a single representative site in each group. In addition, it is found that a common forecasting model can be successfully used for a number of different groups, which saves the operator's workload without deteriorating the forecasting accuracy.
The optimal generation schedule of controllable generators is necessary considering both the supply-demand balancing and the economy of generation. The conventional methods minimize the operation cost with a constraint to guarantee that the operating generators can readily change their outputs to meet the demand variation. Installation of nature-affected generations such as wind turbines and photovoltaic generation makes it harder for operating generators to compensate the variation, because their forecasts contain significant errors. In this research, the damage caused by possible supply-demand imbalance arising from the large uncertainties of forecasts is evaluated by penalty which represents the necessary additional cost to fix the imbalance. Due to the random nature of the forecast errors, the penalty is evaluated as an expectation. The optimal generation schedule is obtained by minimizing the summation of expectation of penalty and the operation cost. The simulation results indicate that the proposed method can achieve the generation schedule with less total cost than the conventional methods.
This paper presents the design consideration of a low-profile LLC resonant converter using two flat trans- formers. The trend toward high power density, high efficiency, and low profile in power supplies has exposed a number of limitations in the use of magnetic component structures. The LLC resonant converter can be operated at a high switching frequency with high efficiency because the switching loss is reduced by soft-switching. However, flat transformer loss causes problems at a high switching frequency. As a result, temperature of flat transformers becomes high. Therefore, it is necessary to reduce the transformer temperature by analyzing the loss. In this proposed con- verter, flat transformer is integrated into advanced power conversion application systems. Low-profile power module of profile of about 14mm is achieved. Temperature inside of transformer repressed to 66.5 ℃ and an overall efficiency of about 96.5% was obtained.
