Search

To address the heterogeneity problem of sensor data, it is necessary to conduct the Sensor Ontology Matching (SOM) process to find the mappings among diverse sensor data with the same semantics connotation. Currently, many Multi-Objective Evolutionary Algorithms (MOEAs) have been used to match the ontologies, which aim at finding a set of solutions called Pareto Set (PS) in the Pareto Front (PF) to represent a set of trade-off proposals for different Decision Makers (DMs). Being inspired by the success of MOEA with Inverse Model (IM-MOEA) in solving complicated optimization problems, in this work, an Improved IM-MOEA (I-IM-MOEA)-based matching technique is further proposed to enhance the algorithm’s matching efficiency as well as the alignment’s quality. To overcome the drawback of IM-MOEA that has poor performance on irregular PF, an adjusted selection mechanism is employed to avert the massive reduction in non-domination solutions on irregular PF, a dynamic Reference Vectors (RVs) is used to decrease the computational resources and boost the efficiency of the algorithm, and a local search strategy is introduced to promote the results’ quality. The experiment employs the benchmark provided by Ontology Alignment Evaluation Initiative (OAEI) and three sensor ontologies to assess the performance of I-IM-MOEA, and the experimental results show that I-IM-MOEA is both effective and efficient.

Location-based services are widely used in mobile applications, which not only bring convenience, but also cause serious privacy concerns. Based on the characteristics of social network, this work proposes a cooperative protection architecture to model the relationship among users, communities and location-based service. Furthermore, in order to construct K anonymity set, a novel community structure-based trilateral Stackelberg game model is developed for K-anonymity protection. In addition, an optimization method based on the proposed model is designed by the backward induction process. Finally, the security and the performance under different situations such as the anonymity parameter K and the community structure parameter overlapping weights are analyzed. The analysis results indicate that the proposed model and the optimization method are effective for privacy protection and can achieve high secure in location-based services.

Traffic forecasting remains an active area of research in the transport and data science fields. Decision-makers rely on traffic forecasting models for both policy-making and operational management of transport facilities. The wealth of spatial and temporal real-time data increasingly available from traffic sensors on roads provides a valuable source of information for policymakers. This paper adopts the Long Short-Term Memory (LSTM) recurrent neural network to predict speed by considering both the spatial and temporal characteristics of real-time sensor data. A total of 288,653 real-life traffic measurements were collected from detector stations on the Eastern Freeway in Melbourne/Australia. A comparative performance analysis among different models such as the Recurrent Neural Network (RNN) that has an internal memory that is able to remember its inputs and Deep Learning Backpropagation (DLBP) neural network approaches are also reported. The LSTM results showed average accuracies in the outbound direction ranging between 88 and 99 percent over prediction horizons between 5 and 60 min, and average accuracies between 96 and 98 percent in the inbound direction. The models also showed resilience in accuracies as the prediction horizons increased spatially for distances up to 15 km, providing a remarkable performance compared to other models tested. These results demonstrate the superior performance of LSTM models in capturing the spatial and temporal traffic dynamics, providing decision-makers with robust models to plan and manage transport facilities more effectively.

Literature Based Discovery (LBD) aims at building bridges between existing literatures and discovering new knowledge from them. Biomedical ontology is such a literature that provides an explicit specification on biomedical knowledge, i.e., the formal specification of the biomedical concepts and data, and the relationships between them. However, since biomedical ontologies are developed and maintained by different communities, the same biomedical information or knowledge could be defined with different terminologies or in different context, which makes the integration of them becomes a challenging problem. Biomedical ontology matching can determine the semantically identical biomedical concepts in different biomedical ontologies, which is regarded as an effective methodology to bridge the semantic gap between two biomedical ontologies. Currently, Evolutionary Algorithm (EA) is emerging as a good methodology for optimizing the ontology alignment. However, EA requires huge memory consumption and long runtime, which make EA-based matcher unable to efficiently match biomedical ontologies. To overcome these problems, in this paper, we define a discrete optimal model for biomedical ontology matching problem, and utilize a compact version of Evolutionary Algorithm (CEA) to solve it. In particular, CEA makes use of a Probability Vector (PV) to represent the population to save the memory consumption, and introduces a local search strategy to improve the algorithm’s search performance. The experiment exploits Anatomy track, Large Biomed track and Disease and Phenotype track provided by the Ontology Alignment Evaluation Initiative (OAEI) to test our proposal’s performance. The experimental results show that CEA-based approach can effectively reduce the runtime and memory consumption of EA-based matcher, and determine high-quality biomedical ontology alignments.

Evolutionary multi-task optimization (EMTO) is a newly emerging research area which studies on how to solve multiple optimization problems simultaneously using evolutionary algorithms (EAs) so that useful knowledge (e.g. promising candidate solutions) obtained when solving one task can be transferred and reused to facilitate solving some other tasks. In EMTO, how to effectively transfer and reuse knowledge among multiple tasks is an important subject of study. Among existing EMTO techniques, DE-based EMTO algorithms deserve special attention because DE, as one of the most popular EAs, has achieved remarkable feats for solving challenging (single-task) optimization problems in the past although DE-based EMTO is not yet extensively studied. In this paper, we propose a general multitasking DE (MTDE) framework which aims to help more clearly understand the focuses and differences of existing and potential works on DE-based EMTO. We implement this framework via three commonly-used DE search schemes and perform a thorough empirical study on two DE-specific knowledge reuse strategies which are based upon base and differential vectors, respectively. Experiments on 18 MTO demonstrate that the base vector based knowledge reuse strategy outperforms the differential vector based one across all three tested DE search schemes.

Vehicle-to-vehicle communication and probabilistic broadcast are important means for information dissemination in vehicular ad-hoc networks (VANETs). In contrast to static ad-hoc networks, high-speed mobility makes wireless connection between two vehicles intermittent. Nevertheless, some kinds of information do not rely on continuous connection for transmission. An important fact is that numerous vehicles ‘desire’ to have such information as safety-related data and leisure information. In recent years, an excessive number of protocols have attached themselves to critical information dissemination, and the existing protocols already have excellent performance. People severely neglect leisure information because of its low importance. Even so, the dissemination of leisure information is an important issue in VANET when, for example, people take long trips on highways. Therefore, we propose interest-aware probabilistic dissemination (IAPD) of leisure information in VANET, which combines probabilistic broadcast and timer-based broadcast techniques. We also develop a state transition system for leisure-information dissemination. The simulation results show that IAPD cost less than restriction flooding and simple schemes in obtaining a high reception rate.

Stacking storage structure has to be used in profile automated warehouse, which results that the above ones of a rack must to be removed before the rack moving. So the delay of the procedure increases. Aiming at the problem, we firstly defined the mixing-degree of the whole stacking-storage as an important indicator of warehouse wholeness, which can characterize warehouse long-term efficiency. Secondly, a utility function was built to describe the measurement of warehouse status based on the mixing-degree. Finally a scheduling optimization model on constraints of the initialization condition was proposed, with always calculating the warehouse mixing-degree and top statuses. Simulation and field experiment on real Nanping AL. Robots system shows that the proposed algorithm achieves good performance and efficiency in long-term, though no optimization in short-term, which roots from less average expectation of the top-racks moves.

In recent years, researchers have conducted various studies on applying wireless networking technology and mobile devices in education settings. However, research on behavioral patterns in learners' online asynchronous discussions with mobile devices is limited. The purposes of this study are to develop a mobile learning system, mobile interactive teaching feedback system (MITFS), linked to both mobile devices and the internet, to support learners with online asynchronous discussion, and combine content analysis and sequential analysis to compare and contrast the social knowledge construction behavioral patterns of problem-based asynchronous discussion in e-learning and m-learning environments. This study investigated four weeks of online discussions in an ''Introduction to Computer Science'' course involving forty first year university students. The control group (online asynchronous discussion without mobile devices) and the experimental group (online asynchronous discussion with mobile devices) in the group discussions were explored. By using content analysis and sequential analysis for the problem-based online asynchronous discussion of the behavioral patterns and differences between students in control and experimental group, the results showed that using mobile devices in online asynchronous discussion influenced students' learning performance. Some interesting results were found. Firstly, when the students used mobile devices in discussion situations, they could more engage in reflecting thinking, sharing more information, and further facilitating social knowledge construction among group members. Secondly, the experimental group performed better than the control group in terms of participation and diversity in knowledge construction behavioral patterns. Finally, based upon the findings, some implications are proposed for further research.

Today organizations are deeply involved in the Big Data era as the amount of data has been exploding with un-predictable rate and coming from various sources. To process and analyze this massive data, privacy is a major concern together with utility of data. Thus, privacy preservation techniques which target at the balance between utility and privacy begin to be one of the recent trends for big data researchers. In this paper, we discuss a technique for big data privacy preservation by means of clustering method. Here, hierarchical particle swarm optimization (HPSO) is used for clustering similar data. To attain scalability for big data, our method is constructed on the novel cloud infrastructure, MapReduce Hadoop. The method is tested by using a novel UCI dataset and the results are compared with an existing approach.

The present study adopted a differentiated instruction method to non-English majors at a university. An e-learning system was used as a support system for the students to check homework answers and to upload assignments. In addition, an assignment was distributed after a midterm examination to encourage the undergraduates to practice with online materials for learning outside the classroom. Students’ feedback on this teaching method was performed on the e-learning system, which delivered a pre-test and a post-test questionnaire before and at the end of the period during which a differentiated activity sheet was used as an incentive for autonomous self-learning. The results showed that compared with a traditional method in which students received assignments of the same difficulty, the differentiated method, together with supports from the e-learning system, enhanced motivation for students to pursue better scores in English achievement tests.

Modeling human actions in a format that is suitable for computer systems to understand is a target for behavior analysis systems. This work introduces a high level design for a behavioral analysis system based on action sequences. The design is introduced in terms of process modeling. System processes are presented in terms of a set of data flow diagrams (also known as DFDs) of multi levels. They represent the decomposition of all processing components required in such system and the data flows among them. The system is designed to receive structured information for human behaviors and actions and produce insights, predictions and classifications for personal and behavioral characteristics. Proposed process decomposition is introduced as a core step towards design and implementation of a behavior analyzer system.

In this paper, we propose a novel approach based on Compact Genetic Algorithm (CGA) to address the problem of optimizing the aggregation of three different basic similarity measures (Syntactic Measure, Linguistic Measure and Taxonomy-based Measure), and get a single similarity metric in the process of ontology matching. Comparing with conventional Genetic Algorithm (GA), the proposed method is able to dramatically reduce the time and memory consumption while at the same time ensures the correctness and completeness of the alignments. Experiment results show that the proposed approach is effective.

In this paper, a passive continuous authentication system based on both the hard and the soft biometrics is implemented. The passive continuous authentication system keeps verifying the user without interrupting the user concentrating on his work. It also provides the capacity for the machine to recognize who is in front of the terminal, reduces the potential security leak of the user is temporarily absent front the terminal, and denies the invader to login the system with the stolen account and password. Our system forces to logoff the user when the authentication result identifies the user leaves his seat, but the system won't logoff the user if the user only turns his face to some other directions. The reliability of the system is verified by 7 registered users. According to the experimental results, combining the soft and the hard biometrics in our theoretical framework achieves the goal of continuous authentication efficiently and correctly.

This chapter reviews the basic idea and processes in data mining and some algorithms within the field of evolutionary computing. The authors focus on introducing the algorithms of computational intelligence since they are useful tools for solving problems of optimization, data mining, and many kinds of industrial issues. A feasible model of combining computational intelligence with data mining is presented at the end of the chapter with the conclusions.

In this paper, we proposed a novel optimization model, which combines bacterial foraging with genetic algorithm. Though these two well-known optimization algorithms have their own good points, they also have their own drawbacks respectively. In our work, a combined evolutional model, bacterial-GA foraging, is proposed. Via applying this new model, experimental results indicate that the new combined model performs much better performance than applying any of these two algorithms singly.