In this paper, we propose (a) fuzzy multiple objective linear programming models for the Supplier Selection and Order Allocation (SSOA) problem under fuzzy demand and volume/quantity discount environments, and (b) an analysis of how to select the suitable aggregation operator based on the risk preferences of decision makers. The aggregation operators under consideration are additive, maximin, and augmented operators while the risk preferences are classified as risk-averse, risk-taking, and risk-neutral ones. The suitabilities of aggregation operators and risk preferences of decision makers are analyzed by a statistical technique, considering the average and the lowest satisfaction levels of the supplier selection criteria, based on numerical examples. Analysis results reveal that decision makers with different risk preferences will prefer only some aggregation operators and models. Moreover, a particular aggregation operator and model may generate a dominated solution for some situations. Thus, it should be applied with caution.

Textual product reviews posted by previous shoppers have been serving as an important source of information that helps on-line shoppers to make their decisions. However, reading through all the reviews of a product is usually a time-demanding and frustrating task, especially when those reviews deliver conflicting information. Therefore, it is of great practical value to develop techniques to automatically generate brief but accurate summaries for the numerous reviews on shopping websites. There are currently two main research streams in review mining: one is joint aspect discovery and sentiment classification, the other one is aspect-level ratings and weights approximation. There exist a number of models in each of the two areas. However, no previous work that aims to solve the two problems simultaneously has been proposed. In this paper we propose Rating Supervised Latent Topic Model to integrate the two problems into an unified optimisation problem. In the proposed model, we employ a latent topic model for aspect discovery and sentiment classification and use a regression model to approximate aspect-level ratings and weights based on the output of the topic model. We test the proposed model on a review dataset crawled from Amazon.com. The preliminary experiment results show that the proposed model outperforms a number of state-of-the-art models by a considerable margin.

The notions of k-monotonicity and superadditivity for non-additive measures (e.g., capacity and cooperative games) are used as indices to measure the complementarity of criteria/coalitions in decision-making involving multiple criteria and/or cooperative game theory. To avoid exponential complexity in capacity-based multicriteria decision-making models, k-additive capacities and/or 𝒞-decomposable capacities are often adopted. While, in cooperative game theory, under communication-restricted situations, some coalitions cannot generally be formed. This paper investigates the inheritance of complementary relationships/effects in non-additive measures with restricted domains (or under bounded interactions).

In this paper, a clustering algorithm for relational data based on q-divergence between memberships and variables that control cluster sizes is proposed. A conventional method for vectorial data is first presented for interpretation as the regularization of another conventional method with q-divergence. With this interpretation, a clustering algorithm for relational data, based on q-divergence, is then derived from an optimization problem built by regularizing the conventional method with q-divergence. A theoretical discussion reveals the property of the proposed method. Numerical results are presented that substantiate this property and show that the proposed method outperforms two conventional methods in terms of accuracy.

Wireless mesh networks (WMNs) are the only cost-effective networks that support seamless connectivity, wide area network (WAN) coverage, and mobility features. However, the rapid increase in the number of users on these networks has brought an upsurge in competition for available resources and services. Consequently, factors such as link congestion, data collisions, link interferences, etc. are likely to occur during service discovery on these networks. This further degrades their quality of service (QoS). Therefore, the quick and timely discovery of these services becomes an essential parameter in optimizing the performance of service discovery on WMNs. In this paper, we present the design and implementation of an enhanced service discovery model that solves the performance bottleneck incurred by service discovery on WMNs. The proposed model integrates the particle swarm optimization (PSO) and ant colony optimization (ACO) algorithms to improve QoS. We use the PSO algorithm to assign different priorities to services on the network. On the other hand, we use the ACO algorithm to effectively establish the most cost-effective path whenever each transmitter has to be searched to identify whether it possesses the requested service(s). Furthermore, we design and implement the link congestion reduction (LCR) algorithm to define the number of service receivers to be granted access to services simultaneously. We simulate, test, and evaluate the proposed model in Network Simulator 2 (NS2), against ant colony-based multi constraints, QoS-aware service selection (QSS), and FLEXIble Mesh Service Discovery (FLEXI-MSD) models. The results show an average service discovery throughput of 80%, service availability of 96%, service discovery delay of 1.8 s, and success probability of service selection of 89%.

To handle a large-scale object, a two-stage clustering method has been previously proposed. The method generates a large number of clusters during the first stage and merges clusters during the second stage. In this paper, a novel two-stage clustering method is proposed by introducing cluster validity measures as the merging criterion during the second stage. The significant cluster validity measures used to evaluate cluster partitions and determine the suitable number of clusters act as the criteria for merging clusters. The performance of the proposed method based on six typical indices is compared with eight artificial datasets. These experiments show that a trace of the fuzzy covariance matrix W_tr and its kernelization KW_tr are quite effective when applying the proposed method, and obtain better results than the other indices.

A clustering method referred to as K-member clustering classifies a dataset into certain clusters, the size of which is more than a given constant K. Even-sized clustering, which classifies a dataset into even-sized clusters, is also considered along with K-member clustering. In our previous study, we proposed Even-sized Clustering Based on Optimization (ECBO) to output adequate results by formulating an even-sized clustering problem as linear programming. The simplex method is used to calculate the belongingness of each object to clusters in ECBO. In this study, ECBO is extended by introducing ideas that were introduced in K-means or fuzzy c-means to resolve problems of initial-value dependence, robustness against outliers, calculation costs, and nonlinear boundaries of clusters. We also reconsider the relation between the dataset size, the cluster number, and K in ECBO. Moreover, we verify the effectiveness of the variants of ECBO based on experimental results using synthetic datasets and a benchmark dataset.

Compressive sensing (CS) of high-order data such as hyperspectral images, medical imaging, video sequences, and multi-sensor networks is certainly a hot issue after the emergence of tensor decomposition. Actually, the reconstruction accuracy with current algorithms is not ideal in some cases of noise. In this paper, we propose a new method that can recover noisy 3-D images from a reduced set of compressive measurements. First, multi-way compressive measurements are performed using Gaussian random matrices. Second, the mapping relationship between the variance of noise and the reconstruction threshold is found. Finally, the original images are recovered through reconstruction of pseudo inverse based on threshold selection. We experimentally demonstrate that the proposed method outperforms other similar methods in both reconstruction accuracy (within a range of the compression ratios and different variances of noise) and processing speed.

Availability of wind speed information is of great importance for maximization of wind energy extraction in wind energy conversion systems. The wind speed is commonly obtained from a direct measurement employing a number of anemometers installed surrounding the wind turbine. In this paper a sensorless fuzzy wind speed estimator is proposed. The estimator is easy to build without any training or optimization. It works based on the fuzzy logic principles heuristically inferred from the typical wind turbine power curve. The wind speed estimation using the proposed estimator was simulated during the operation of a squirrel-cage induction generator-based wind energy conversion system. The performance of the proposed estimator was verified by the well estimated wind speed obtained under the wind speed variation.

Lower extremity fatigue is a risk factor for falls and injuries. This paper proposes a machine learning system to detect fatigue states, which considers the different influences of common daily activities on physical health. A wearable inertial unit is devised for gait data acquisition. The collected data are reorganized into nine data subsets for dimension reduction, and then preprocessed via gait cycle division, visualization, and oversampling. Then, a heterogeneous ensemble learning voting method is employed to train nine classifiers. The results indicate that the method reaches an accuracy of 92%, which is obtained by the plurality voting method using data subset prediction classes. Comparing the results shows that the final result is more accurate than the results of each individual data subset, and the heterogeneous voting method is advantageous when balancing out individual weaknesses of a set of equally well-performing models.

We propose a Healthy Eating Habits Support System (HEHSS) which considers user taste preferences and nutritional balance. The proposed system comprises a Nutritional Management System (NMS) and a Kansei Retrieval System (KRS). The NMS generates nutritionally balanced menus using the tabu search method. The KRS learns user taste preferences through interaction with a user, and then uses this information to recommend appropriate menus for that user from those generated by the NMS. Consequently, the HEHSS recommends menus that consider nutritional balance and match the user’s taste preferences. Simulation results demonstrate that the HEHSS recommended menus that satisfied nutritional needs and learned a user’s taste preferences with greater than 80% accuracy when we focused on liked and disliked tastes after continuous use for a long period (approximately 2 months).

This paper proposes a class of dynamic P systems with constraint of discrete Morse function (DMD P systems). Membrane structure is extended on complex. Rules control activities of membranes. New classes of rules and mechanism to change types of rules by discrete gradient vector field are provided as well. DMD P system extends P systems both in structures and rules. Solving air quality evaluation problem in linear time verifies the effectiveness of DMD P systems. Since air quality evaluation problem has significance in many areas. The new P systems provide an alternative for traditional membrane computing.

Implanted knee kinematics recognition is required in total knee arthroplasty (TKA), which replaces damaged knee joint with artificial one. The 3-D kinematics of implanted knee in-vivo is used to quantify the knee function for diagnosis of TKA patients and to evaluate the design of TKA prosthesis and surgical techniques. There are some methods for the implanted knee kinematics estimation, however, those methods are classified into still image analysis. The discontinuous knee kinematics estimated by the still image analysis is not considered as the actual knee kinematics. This paper proposes an kinematics recognition method for implanted knee using particle filter. The proposed method estimates the 3-D pose/position parameters, which are varying in time, based on a priori knowledge of time evolution of the parameters represented by random walk models and utilizing similarity between acquired DR image frame and synthesized DR image based on hypothesized value of the parameters. The experimental results showed that the proposed method successfully estimated the 3-D implanted knee kinematics with an accuracy of 1.61 mm and 0.32°.

White canes are widely used as tools to assist visually impaired persons to walk. The white cane is used to acquire environmental information as a clue for walking. It is important to know what information can be acquired by using the cane, how accurate the acquired information is, and how we can improve the information and accuracy. Clarification of these questions would contribute to development of white canes that enable better acquisition of environmental information. It would also assist in the design of buildings with appropriate guidance for visually impaired persons on the basis of information acquired through the canes. In this study, in order to acquire basic knowledge about hardness recognition, experiments were performed using four conditions: 1) Tapping or pushing the object with the index finger tip and white cane tip, 2) use or non-use of the auditory sense to study its usefulness, 3) different ways of grasping the cane, and 4) different number of checks. Nine visually impaired persons who usually walked alone using white canes participated in the experiments. They estimated the hardness of a rubber sheet under various combinations of the operation conditions and hearing conditions. Results showed that the number of checks had little effect on the user’s estimation of the hardness of the sheet. For the recognition of the contact target hardness using the white cane, it was effective to simultaneously use information from different modalities, namely tactile and auditory information. We also observed that, when pushing the cane with the index fingers, the users could feel the objects as if they had directly touched them with the fingers.

Wind generation is one of the fastest growing resources among renewable energies worldwide including Japan. As Japan is an island country surrounded by ocean, the on-shore landscape topography suitable for wind generation is limited. Therefore, based on the wind map until the year 2030, it is expected that new off-shore wind generation installations will be more suitable. For this reason, it is very important to determine the wind characteristics of the candidate areas for installing wind generation; however, in most off-shore installation sites, availability of weather condition data is poor and significant time and cost are required to accurately measure pin-point wind/weather conditions data. In this study, our goal is to project the wind speed of an unseen area (where weather condition data are not available) by mapping the seen areas (where weather condition data are available) around the target area using a modularized Artificial Neural Network (ANN) referred to as a Self-Organization Map (SOM). By learning the correlation between modularized ANNs of seen and unseen areas, the result of this temporal and spatial projection is the prediction of wind speed of a target area. We believe that the proposed technique will significantly reduce the amount of time and cost involved in selection of off-shore installation sites. Moreover, it should contribute to accelerated development and implementation of off-shore wind power generation in the future.

This study proposes an iterative closest shape point (ICSP) registration method based on regional shape maps for 3D face recognition. A neutral expression image randomly selected from a face database is considered as the reference face. The point-to-point correspondences between the input face and the reference face are achieved by constructing the points’ regional shape maps. The distance between corresponding point pairs is then minimized by iterating through the correspondence findings and coordinate transformations. The vectors composed of the closest shape points obtained in the last iteration are regarded as the feature vectors of the input face. These 3D face feature vectors are finally used for both training and recognition using the Fisherface method. Experiments are conducted using the 3D face database maintained by the Chinese Academy of Science Institute of Automation (CASIA). The results show that the proposed method can effectively improve 3D face recognition performance.

This paper identifies four variables to reveal the internal mechanisms of the entropy measures on intuitionistic fuzzy sets (IFSs) and interval-valued intuitionistic fuzzy sets (IVIFSs). First, four variables are used to propose a pair of generalized entropy measures on IFSs and IVIFSs. Second, three specific entropy measures are put forward to illustrate the effectiveness of the generalized entropy measure functions. Third, a novel multiple attribute decision-making approach under an intuitionistic fuzzy environment is proposed. The superiority of the decision-making approach is that the weight values of the attributes are obtained by their related entropy measures. Finally, the performance of the proposed entropy regulations on IFSs and IVIFSs is illustrated through a mode assessment example on open communities.