Tiranee Achalakul

The best-so-far selection in Artificial Bee Colony algorithm

The Artificial Bee Colony (ABC) algorithm is inspired by the behavior of honey bees. The algorithm is one of the Swarm Intelligence algorithms explored in recent literature. ABC is an optimization technique, which is used in finding the best solution from all feasible solutions. However, ABC can sometimes be slow to converge. In order to improve the algorithm performance, we present a modified method for solution update of the onlooker bees in this paper. In our method, the best feasible solutions found so far are shared globally among the entire population. Thus, the new candidate solutions are more likely to be close to the current best solution. In other words, we bias the solution direction toward the best-so-far position. Moreover, in each iteration, we adjust the radius of the search for new candidates using a larger radius earlier in the search process and then reduce the radius as the process comes closer to converging. Finally, we use a more robust calculation to determine and compare the quality of alternative solutions. We empirically assess the performance of our proposed method on two sets of problems: numerical benchmark functions and image registration applications. The results demonstrate that the proposed method is able to produce higher quality solutions with faster convergence than either the original ABC or the current state-of-the-art ABC-based algorithm.

A distributed spectral-screening PCT algorithm

This paper describes a novel distributed algorithm for use in remote-sensing, medical image analysis, and surveillance applications. The algorithm combines spectral-screening classification with the principal component transform, and human-centered mapping. It fuses a multi- or hyper-spectral image set into a single color-composite image that maximizes the impact of spectral variation on the human visual system. The algorithm operates on distributed collections of shared-memory multiprocessors that are connected through high-performance networking. Scenes taken from a standard 210 frame remote-sensing data set, collected with the hyper-spectral digital imagery collection experiment airborne imaging spectrometer, are used to assess the algorithms image quality, performance, and scaling. The algorithm is supported with a predictive analytical model that allows its performance to be assessed for a wide variety of typical variations in use. For example, changes to the number of spectra, image resolution, processor speed, memory size, network bandwidth/latency, and granularity of decomposition. The motivation in building a performance model is to assess the impact of changes in technology and problem size associated with different applications, allowing cost-performance tradeoffs to be assessed.

ABC-GSX: A hybrid method for solving the Traveling Salesman Problem

An optimization problem is a problem of finding the best solution from all possible solutions. In most computer science and mathematical applications, the decision to select the best solution is not polynomially bounded. Heuristics approaches are thus often considered to solve such NP-hard problems. In our work, we focus on developing a heuristic method to solve a combinatorial optimization problem known as the Traveling Salesman Problem or TSP. Our technique implements the Artificial Bee Colony algorithm, which is inspired by the decision making process of the honey bees in finding optimal food sources. We extend the ABC algorithm with Greedy Subtour Crossover to improve the precision. In this hybrid procedure, the exploitation process in the ABC algorithm is improved upon by the Greedy Subtour Crossover method. The new proposed method is called ABC-GSX. We then empirically assess performance of our proposed work using functions from a standard TSP library. Experimental results show improvements in both precision and computational time compared to techniques presented in recent literatures.

The best-so-far ABC with multiple patrilines for clustering problems

Clustering is an important process in many application domains such as machine learning, data mining, pattern recognition, image analysis, information retrieval, and bioinformatics. The main objective of clustering is to search for hidden patterns that may exist in datasets. Since the clustering problem is considered to be NP-hard, previous research has applied bio-inspired heuristic methods to solve such problems. In this paper we propose an effective method for clustering using an algorithm inspired by the decision making processes of bee swarms. The algorithm is called the Best-so-far Artificial Bee Colony with multiple patrilines. In the Best-so-far method, the solution direction is biased toward the Best-so-far solution rather than a neighboring solution proposed in the original Artificial Bee Colony algorithm. We introduce another bee-inspired concept called multiple patrilines to further improve the diversity of solutions and allow the calculations to be distributed among multiple computing units. We empirically assess the performance of our proposed method on several standard datasets taken from the UCI Machine Learning Repository. The results show that the proposed method produces solutions that are as good as or better than the current state-of-the-art clustering techniques reported in the literature. Furthermore, to demonstrate the computing performance and scalability of the algorithm, we assess the algorithm on a large disk drive manufacturing dataset. The results indicate that our distributed Best-so-far approach is scalable and produces good solutions while significantly improving the processing time.

Artificial bee colony algorithm on distributed environments

Artificial Bee Colony (ABC) is a metaheuristic approach in which a colony of artificial bees cooperates in finding good solutions for numerical optimization problems. ABC is adopted widely for use in several domains of solution optimization. However, the algorithm generally requires a considerably large computational time and resources. In order to enhance the performance of this algorithm for a large problem size, we introduce a distributed version of ABC. In our parallel algorithm, the entire bee colony is decomposed into several subgroups. Each subgroup then performs a local search concurrently on each processor node. The local best solutions are then exchanged among processor nodes. The algorithm implementation utilizes the message passing technique as a communication paradigm. We then empirically assess the performance based on both result accuracy and algorithms efficiency. The experimental results show improvement in both solution quality and computing time when comparing to the sequential ABC algorithm.

Cost optimal scheduling in IaaS for dependent workload with particle swarm optimization

Optimizing cloud provisioning for scientific workflow applications is a challenging problem, since the workflows generally contain dependency between tasks and require specific deadlines. Usually, cloud providers offer many options to the consumers. These options include the number of virtual machines, the type of each virtual machine and the purchasing method for each machine. Currently, cloud provisioning cost optimization is an active research topic. Most of this literature is concerned with task scheduling, cloud option selection, and cloud option selection for scientific workflow applications. However, research that attempts to find solutions which cover both cloud option selection and workflow task scheduling is very limited. In this paper, we focus on optimizing the cost of purchasing infrastructure-as-a-service cloud capabilities to achieve scientific work flow execution within the specific deadlines. The proposed system considers the number of purchased instances, instance types, purchasing options, and task scheduling as constraints in an optimization process. Particle swarm optimization augmented with a variable neighborhood search technique is used to find the optimal solution. Our approach finds the configurations of purchasing options with the optimum budget for a specified workflow application based on the required performance. The solutions from the proposed system show promising performance from the perspectives of the total cost and fitness convergence when compared with other state-of-the-art algorithms.

Reducing bioinformatics data dimension with ABC-kNN

Abstract Analyzing a large amount of data often consumes extensive computational resources and execution time. However, sometime all data features do not equally contribute to the end results. Thus, it is plausible to identify the major contributing features and use them as representatives of the data. Other features with low contribution can be eliminated to reduce the time/resource consumption in data analysis. One of the promising application domains for such a feature selection process is bioinformatics. The need for dimension reduction, which is the process to reduce unnecessary features from the original data, arises because biological data can be massive, with tens of thousands of features to be explored. The objective of this study is to design an effective algorithm that can selectively remove irrelevant dimensions from data describing complex biological processes while preserving the semantics of the original data. This research proposes the adoption of the Artificial Bee Colony (ABC) as a novel method for data dimension reduction in classification problems. ABC, an efficient heuristic method based on swarm intelligence, is used to select the optimal subset of dimensions from the original high-dimensional data while retaining a subset that satisfies the defined objective. The k-Nearest Neighbor (kNN) method is then used for fitness evaluation within the ABC framework. In this research, ABC and kNN have been modified and bundled together to create an effective dimension reduction method. The proposed algorithm is validated in two distinct application domains: Gene expression analysis, and autistic behaviors study. The experimental results exhibit good solution quality as well as good computational performance.

Real-time multi-spectral image fusion

This paper describes a novel real‐time multi‐spectral imaging capability for surveillance applications. The capability combines a new high‐performance multi‐spectral camera system with a distributed algorithm that computes a spectral‐screening principal component transform (PCT). The camera system uses a novel filter wheel design together with a high‐bandwidth CCD camera to allow image cubes to be delivered at 110 frames

A multiple-objective workflow scheduling framework for cloud data analytics

^{-1}

Failure Prediction of Data Centers Using Time Series and Fault Tree Analysis

s with a spectral coverage between 400 and 1000 nm. The filters used in a particular application are selected to highlight a particular object based on its spectral signature. The distributed algorithm allows image streams from a dispersed collection of cameras to be disseminated, viewed, and interpreted by a distributed group of analysts in real‐time. It operates on networks of commercial‐off‐the‐shelf multiprocessors connected with high‐performance (e.g. gigabit) networking, taking advantage of multi‐threading where appropriate. The algorithm uses a concurrent formulation of the PCT to de‐correlate and compress a multi‐spectral image cube. Spectral screening is used to give features that occur infrequently (e.g. mechanized vehicles in a forest) equal importance to those that occur frequently (e.g. trees in the forest). A human‐centered color‐mapping scheme is used to maximize the impact of spectral contrast on the human visual system.