Variable space diversity, crossover and mutation in MOEA solving many-objective knapsack problems

In this work, we analyze variable space diversity of Pareto optimal solutions (POS) and study the effectiveness of crossover and mutation operators in evolutionary many-objective optimization. First we examine the diversity of variables in the true POS on many-objective 0/1 knapsack problems with up to 20 items (bits), showing that variables in POS become noticeably diverse as we increase the number of objectives. We also verify the effectiveness of conventional two-point and uniform crossovers, Local Recombination that selects mating parents based on proximity in objective space, and two-point and uniform crossover operators which Controls the maximum number of Crossed Genes (CCG). We use NSGA-II, SPEA2, IBEA _??+? and MSOPS, which adopt different selection methods, and many-objective 0/1 knapsack problems with $n=\{100,250,500,750,\mbox{1,000}\}$ items (bits) and m?=?{2,4,6,8,10} objectives to verify the search performance of each crossover operator. Simulation results reveal that Local Recombination and CCG operators significantly improve search performance especially for NSGA-II and MSOPS, which have high diversity of genes in the population. Also, results show that CCG operators achieve higher search performance than Local Recombination for m?≥?4 objectives and that their effectiveness becomes larger as the number of objectives m increases. In addition, the contribution of CCG and mutation operators for the solutions search is analyzed and discussed.     

A content search system considering the activity and context of a mobile user

People routinely carry mobile devices in their daily lives and obtain a variety of information from the Internet in many different situations. In searching for information (content) with a mobile device, a user’s activity (e.g., moving or stationary) and context (e.g., commuting in the morning or going downtown in the evening) often change, and such changes can affect the user’s degree of concentration on his or her mobile device’s display and information needs. Therefore, a search system should provide the user with an amount of information suitable for the current activity and a type of information suitable for the current context. In this study, we present the design and implementation of a content search system that considers a mobile user’s activity and context, with the goal of reducing the user’s operation load for content search. The proposed system switches between two kinds of content search systems according to the user’s activity: the location-based content search system is activated when the user is stationary (e.g., standing and sitting), while a menu-based content search system is activated when the user is moving (e.g., walking). Both systems present information according to user context. The location-based system presents detailed information via menus and a map according to location-based categories. The menu-based system presents only a few options to enable users to get content easily. Through user experiments, we confirmed that participants could get desired information more easily with this system than with a commercial search system.     

Some properties of mullite powders prepared by chemical vapour deposition

The sinterability of mullite (3Al₂O₃·2SiO₂) powder prepared by chemical vapour deposition was examined to improve the conditions for fabricating dense mullite ceramics. The starting powder contained not only mullite, but also a small amount of -Al₂O₃ (Al-Si spinel) and amorphous material. Although the compressed powder was fired at a temperature between 1550 and 1700 °C for 1, 3 and 5 h, the relative densities of the sintered compacts were limited to 90%: (i) due to the creation of pores/microcracks during the solid state reaction (1100–1350 °C), and (ii) due to restriction on the rearrangement of grains because the amount of liquid phase (1550–1700 °C) was insufficient. Calcination of the starting powder was effective for preparation of easily sinterable powder with homogeneous composition. When the compact formed by compressing the calcined powder at 1400 °C for 1 h was fired at 1650 °C for 3 h, the relative density was raised up to 97.2%; moreover, mullite was the only phase detected from the sintered compact. The sintered compact was composed of polyhedral grains with sizes of 1–2 m and elongated grains with long axes of 6 m.     

Effects of fiber orientation angles of fiber-reinforced plastic on sand solid particle erosion behaviors

Solid particle erosion in industrial applications has been a serious problem in many engineering fields. Earlier studies on fiber-reinforced plastic (FRP) composites were mainly focusing on the erosive wear behavior at several different impact angles. However, the effect of fiber orientation on FRP composites has not been thoroughly investigated. Since fiber orientation is one of the important factors in which causing erosive wear damages to FRP composites, in order to understand the virtue of this problem, it is important to investigate the effect of fiber orientation at different impact angles. In this research, the effect of fiber orientation of unidirectional fiber-reinforced plastic composites on erosive wear behavior was studied. Sandblasting-type erosion tests were conducted on the FRP composites with fiber orientation ranging at three impact angles to clarify the relation between fiber orientation and erosive wear behavior. The Dyneema fiber (ductile material) and the carbon fiber (brittle material) were used for the reinforcement fiber in FRP. From the result, it is confirmed that CFRP composites with higher fiber orientation angle erode faster than the composites with lower fiber orientation angle. But the erosion characteristic of DFRP was almost the same regardless of the fiber orientation angle. The damaged surfaces of the FRP composites were then analyzed using scanning electron microscopy and the possible erosion wear mechanisms were investigated.     

The effect of lutein supplementation on visual fatigue: A psychophysiological analysis

We used psychophysiological technology to examine the effect of an oral supplement, a combination of lutein, zeaxanthin and blackcurrant extract (LUT), on visual fatigue, within the context of a randomized, double-blind, placebo-controlled cross-over trial. The LUT supplement and placebo samples were randomly assigned to thirteen participants, who took the samples for two LUT (and vice versa) for another 2 week. Each participant completed visual proof reading tasks for 2 h during each of four testing sessions. Saccade tests were administered before and after the proof reading task, during which the participants moved their eyes back and forth between two targets positioned in the center of two checkerboards. We recorded EEG, EOG, heart rate, and facial muscle potential/performance during the saccade tests. Blood pressure was measured and subjective fatigue and stress scores were collected before and after the proof reading task. We averaged EEG starting at saccade offset in order to analyze eye fixation related potentials (EFRP). Our results suggested that the proof reading task induced visual fatigue. An analysis of EFRP and other psychophysiological data revealed significant differences between the LUT and placebo conditions. These results suggest that supplementation with LUT could help to reduce symptoms of visual fatigue.     

Parsing human skeletons in an operating room

Multiple human pose estimation is an important yet challenging problem. In an operating room (OR) environment, the 3D body poses of surgeons and medical staff can provide important clues for surgical workflow analysis. For that purpose, we propose an algorithm for localizing and recovering body poses of multiple human in an OR environment under a multi-camera setup. Our model builds on 3D Pictorial Structures and 2D body part localization across all camera views, using convolutional neural networks (ConvNets). To evaluate our algorithm, we introduce a dataset captured in a real OR environment. Our dataset is unique, challenging and publicly available with annotated ground truths. Our proposed algorithm yields to promising pose estimation results on this dataset.     

Audio-visual speech recognition integrating 3D lip information obtained from the Kinect

Audio-visual speech recognition (AVSR) has shown impressive improvements over audio-only speech recognition in the presence of acoustic noise. However, the problems of region-of-interest detection and feature extraction may influence the recognition performance due to the visual speech information obtained typically from planar video data. In this paper, we deviate from the traditional visual speech information and propose an AVSR system integrating 3D lip information. The Microsoft Kinect multi-sensory device was adopted for data collection. The different feature extraction and selection algorithms were applied to planar images and 3D lip information, so as to fuse the planar images and 3D lip feature into the visual-3D lip joint feature. For automatic speech recognition (ASR), the fusion methods were investigated and the audio-visual speech information was integrated into a state-synchronous two stream Hidden Markov Model. The experimental results demonstrated that our AVSR system integrating 3D lip information improved the recognition performance of traditional ASR and AVSR system in acoustic noise environments.     

Platform design for large-scale artificial market simulation and preliminary evaluation on the K computer

Artificial market simulations have the potential to be a strong tool for studying rapid and large market fluctuations and designing financial regulations. High-frequency traders, that exchange multiple assets simultaneously within a millisecond, are said to be a cause of rapid and large market fluctuations. For such a large-scale problem, this paper proposes a software or computing platform for large-scale and high-frequency artificial market simulations (Plham: /pl\(\Lambda\)m). The computing platform, Plham, enables modeling financial markets composed of various brands of assets and a large number of agents trading on a short timescale. The design feature of Plham is the separation of artificial market models (simulation models) from their execution (execution models). This allows users to define their simulation models without parallel computing expertise and to choose one of the execution models they need. This computing platform provides a prototype execution model for parallel simulations, which exploits the variety in trading frequency among traders, that is, the fact that some traders do not require up-to-date information of markets changing in millisecond order. We evaluated a prototype implementation on the K computer using up to 256 computing nodes.     

Path-planning and navigation of a mobile robot as discrete optimization problems

There is huge diversity among navigation and path-planning problems in the real world because of the enormous number and great variety of assumptions about the environments, constraints, and tasks imposed on a robot. To deal with this diversity, we propose a new solution to the path-planning and navigation of a mobile robot. In our approach, we formulated the following two problems at each time-step as discrete optimization problems: (1) estimation of a robot's location, and (2) action decision. For the first problem, we minimize an objective function that includes a data term, a constraint term, and a prediction term. This approach is an approximation of Markov localization. For the second problem, we define and minimize another objective function that includes a goal term, a smoothness term, and a collision term. Simulation results show the effectiveness of our approach. This work was presented in part at the Fifth International Symposium on Artificial Life and Robotics, Oita, Japan, January 26–28, 2000     

Minimizing False Positives of a Decision Tree Classifier for Intrusion Detection on the Internet

Machine learning or data mining technologies are often used in network intrusion detection systems. An intrusion detection system based on machine learning utilizes a classifier to infer the current state from the observed traffic attributes. The problem with learning-based intrusion detection is that it leads to false positives and so incurs unnecessary additional operation costs. This paper investigates a method to decrease the false positives generated by an intrusion detection system that employs a decision tree as its classifier. The paper first points out that the information-gain criterion used in previous studies to select the attributes in the tree-constructing algorithm is not effective in achieving low false positive rates. Instead of the information-gain criterion, this paper proposes a new function that evaluates the goodness of an attribute by considering the significance of error types. The proposed function can successfully choose an attribute that suppresses false positives from the given attribute set and the effectiveness of using it is confirmed experimentally. This paper also examines the more trivial leaf rewriting approach to benchmark the proposed method. The comparison shows that the proposed attribute evaluation function yields better solutions than the leaf rewriting approach.