Chun Wei Tsai

Big data analytics: a survey

AbstractThe age of big data is now coming. But the traditional data analytics may not be able to handle such large quantities of data. The question that arises now is, how to develop a high performance platform to efficiently analyze big data and how to design an appropriate mining algorithm to find the useful things from big data. To deeply discuss this issue, this paper begins with a brief introduction to data analytics, followed by the discussions of big data analytics. Some important open issues and further research directions will also be presented for the next step of big data analytics.

A hybrid meta-heuristic algorithm for VM scheduling with load balancing in cloud computing

AbstractVirtual machine (VM) scheduling with load balancing in cloud computing aims to assign VMs to suitable servers and balance the resource usage among all of the servers. In an infrastructure-as-a-service framework, there will be dynamic input requests, where the system is in charge of creating VMs without considering what types of tasks run on them. Therefore, scheduling that focuses only on fixed task sets or that requires detailed task information is not suitable for this system. This paper combines ant colony optimization and particle swarm optimization to solve the VM scheduling problem, with the result being known as ant colony optimization with particle swarm (ACOPS). ACOPS uses historical information to predict the workload of new input requests to adapt to dynamic environments without additional task information. ACOPS also rejects requests that cannot be satisfied before scheduling to reduce the computing time of the scheduling procedure. Experimental results indicate that the proposed algorithm can keep the load balance in a dynamic environment and outperform other approaches.

A memetic particle swarm optimization algorithm for solving the DNA fragment assembly problem

Abstract Determining the sequence of a long DNA chain first requires dividing it into subset fragments. The DNA fragment assembly (DFA) approach is then used for reassembling the fragments as an NP-hard problem that is the focus of increasing attention from combinatorial optimization researchers within the computational biology community. Particle swarm optimization (PSO) is one of the most important swarm intelligence meta-heuristic optimization techniques to solve NP-hard combinatorial optimization problems. This paper proposes a memetic PSO algorithm based on two initialization operators and the local search operator for solving the DFA problem by following the overlap–layout–consensus model to maximize the overlapping score measurement. The results, based on 19 coverage DNA fragment datasets, indicate that the PSO algorithm combining tabu search and simulated annealing-based variable neighborhood search local search can achieve the best overlap scores.

Computational awareness for smart grid: A review

Smart grid has been an active research area in recent years because almost all the technologies required to build smart grid are mature enough. It is expected that not only can smart grid reduce electricity consumption, but it can also provide a more reliable and versatile service than the traditional power grid can. Although the infrastructure of smart grid all over the world is far from complete yet, there is no doubt that our daily life will benefit a lot from smart grid. Hence, many researches are aimed to point out the challenges and needs of future smart grid. The question is, how do we use the massive data captured by smart meters to provide services that are as “smart” as possible instead of just automatically reading information from the meters. This paper begins with a discussion of the smart grid before we move on to the basic computational awareness for smart grid. A brief review of data mining and machine learning technologies for smart grid, which are often used for computational awareness, is then given to further explain their potentials. Finally, challenges, potentials, open issues, and future trends of smart grid are addressed.

PSGO: Particle Swarm Gravitation Optimization Algorithm

Particle swarm optimization (PSO) is the most well known of the swarm-based intelligence algorithms. However, the PSO converges prematurely, which rapidly decreases the population diversity, especially when approaching local optima. To improve the diversity of the PSO, we here propose a memetic algorithm called particle swarm gravitation optimization (PSGO). After a specific number of iterations, some individuals selected from the PSO and GSA systems are exchanged by the roulette wheel approach. Finally, to increase the diversities of the PSO and GSA, we introduce a diversity enhancement operator, which is inspired by the crossover operator used in differential evolution algorithms. In evaluations of five benchmark functions, the PSGO significantly outperformed the PSO and Cuckoo search and yielded a superior performance to the GSA of most of instances and computation times. 8

Classification algorithms for interactive multimedia services: a review

Interactive multimedia services, which integrate and unify techniques from a variety of disciplines, have been an active research topic for many years. However, two major challenges need to be overcome to provide a better service: The first is that interactive multimedia systems have to provide the contents a user needs at the right time no matter where the user is located and what device the user is using; the second is that the performance of such systems needs to be improved. Apparently, classification and clustering (also called unsupervised classification) algorithms play an indispensable role in these respects. Thus, this paper contains a review of the classification algorithms for interactive multimedia systems. Also discussed in this paper are several important issues, open questions, and trends.

Big Data Analytics

The age of big data is now coming. But the traditional data analytics may not be able to handle such large quantities of data. The question that arises now is, how to develop a high performance platform to efficiently analyze big data and how to design an appropriate mining algorithm to find the useful things from big data. To deeply discuss this issue, this paper begins with a brief introduction to data analytics, followed by the discussions of big data analytics. Some important open issues and further research directions will also be presented for the next step of big data analytics.

Rectifying the Inconsistent Fuzzy Preference Matrix in AHP Using a Multi-Objective BicriterionAnt

Analytic hierarchy process (AHP) is a decision making tool regarding the criteria analysis to obtain a priority alternative. One of the important issues in comparison matrix of AHP is the consistency. The inconsistent comparison matrix cannot be used to make decision. This paper proposes an algorithm using a modified BicriterionAnt to pursue two objectives intended to rectify the inconsistent fuzzy preference matrix, called MOBAF. The two objectives include minimizing the consistent ratio (CR) and minimizing the deviation matrix, which are in conflict with each other when rectifying the inconsistent matrix. This study uses two pheromone matrices and two heuristic distances matrices to generate the ants tour. To see the performance, MOBAF is implemented to rectify on some inconsistent fuzzy preference matrices. As a result, in addition to being able to rectify the CR, the proposed algorithm also successfully generates some non-dominated solutions that can be considered as optimal solutions.

A memetic gravitation search algorithm for solving DNA fragment assembly problems

The DNA fragment assembly (DFA) problem is among the most critical problems in computational biology. Being NP-hard, it can be efficiently solved via meta-heuristic algorithms, such as the gravitation search algorithm (GSA). GSA is a state-of-the-art swarm-based algorithm particularly suitable for solving NP-hard combinatorial optimization problems. This paper proposes a new memetic GSA algorithm called MGSA. MGSA is a type of overlap-layout-consensus model that is based on tabu search for population initialization. In order to increase the diversity of MGSA, we adapted two operator time-varying maximum velocities in the GSA procedure. Finally we also adapted the simulated annealing-based variable neighborhood search (SA-VNS) to find superior precise solutions. The proposed MGSA algorithm was verified with 19 DNA fragments based on seeking to maximize the overlap score measurements. In comparing the performances of the proposed MGSA and state-of-the-art algorithms, the simulation results demonstrate that the MGSA can achieve the best overlap scores.

Intelligent big data analysis: a review

Big data analysis is definitely an urgent task for most information systems. Its importance and potentials can be found in many recent studies. They buzzed with this research issue because the data we collect and create are increasing at an unprecedented rate. Thus, the data analysis process has to be reconsidered. In this paper, we will first give a brief discussion of big data from different perspectives, such as size of data, characteristics of data, and source of data. Then, data mining and other information retrieval technologies for big data will be addressed. A brief review of other computational intelligence technologies for big data will also be given. Finally, open issues and future research trends using computational intelligence technologies are presented to show their potentials for big data.