An Improvement of Database with Local Search Mechanisms for Genetic Algorithms in Large-Scale Computing Environments

In this study, a new Genetic Algorithm (GA) using the Tabu · Local Search mechanism is proposed. The GA described in this paper is considered a Mega Process GA, which has an effective mechanism to use massive processors, i.e., Mega Processors, in large-scale computing systems. Our proposed method has a GA-specific database that possesses information of searched space and performs a local search for the space that is not searched. Such mechanisms enable us to comprehend the quantitative rate of a searched region during the search. Using this information, the searched space can be expanded linearly as the number of computing resources increases and the exhaustive search is guaranteed under infinite computations. The proposed GA was applied to numerical test functions and the energy minimization problems of protein tertiary structures. The latter problem was performed under a heterogeneous distributed computing environment, which was built up with Grid MP produced by United Devices Inc.KeywordsGenetic AlgorithmLocal SearchExhaustive SearchProtein Tertiary StructureLocal Search PhaseThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Purpose At this work, we propose a local approximation based search method to optimize any function. For this purpose, an approximation method is combined with an estimation filter, and a new local search mechanism is constituted. Design/methodology/approach RBF network is very efficient interpolation method especially if we have sufficient reference data. Here, reference data refers to the exact value of the objective function at some points. Using this capability of RBFs, we can approximately inspect the vicinity each point in search space. Meanwhile, in order to obtain a smooth, rapid and better trajectory toward the global optimum, the alpha-beta filter can be integrated to this mechanism. For better description and visualization, the operations are defined in 2-dimensional search space; but the outlined procedure can be extended to higher dimensions without loss of generality. Findings When compared with our previous studies using conventional heuristic methods, approximation based curvilinear local search mechanism provide better minimization performance for almost all benchmark functions. Moreover computational cost of this method too less than heuristics. The number of iteration down to at least 1/10 compared to conventional heuristic algorithm. Additionally, the solution accuracy is very improved for majority of the test cases. Originality/value This paper proposes a new search approach to solve optimization problems with less cost. For this purpose, a new local curvilinear search mechanism is built using RBF based approximation technique and alpha-beta estimation filter.

Graphics Processing Units (GPUs) have been widely used as accelerators in large-scale heterogeneous computing systems. However, current programming models can only support the utilization of local GPUs. When using non-local GPUs, programmers need to explicitly call API functions for data communication across computing nodes. As such, programming GPUs in large-scale computing systems is more challenging than local GPUs since local and remote GPUs have to be dealt with separately. In this work, we propose a virtual OpenCL (VOCL) framework to support the transparent virtualization of GPUs. This framework, based on the OpenCL programming model, exposes physical GPUs as decoupled virtual resources that can be transparently managed independent of the application execution. To reduce the virtualization overhead, we optimize the GPU memory accesses and kernel launches. We also extend the VOCL framework to support live task migration across physical GPUs to achieve load balance and/or quick system maintenance. Our experiment results indicate that VOCL can greatly simplify the task of programming cluster-based GPUs at a reasonable virtualization cost.

With the parallel computer systems scaling-up, the measure index for performance of the systems demands a shift from traditional "high performance" to "high productivity." This brings a new challenge to defining a synthetic, yet meaningful, measure index of multiple productivity variables; namely computing performance, reliability, energy consumption, parallel software development, etc. Traditional measures for large-scale parallel computer systems merely focus on computing performance, and are incapable of measuring the multiple productivity variables simultaneously in an effective manner. A recently proposed market-related money model, which pursues high utility/cost ratio, relies on money as a measure to consider the multiple productivity variables. Differing from the previous models, this paper proposes a novel system productivity speedup metric for large-scale parallel computer systems. The metric uses speedup instead of money to comprehensively unify the measures of multiple productivity variables. Finally, we propose a trade-off productivity measurement to weigh different productivity variables, to address different design targets. The measurement can facilitate the system evaluation, expose future technique tendencies, and guide future system design.

This tutorial presents a novel search system—the Attractor-Based Search System (ABSS)—that can solve the Traveling Salesman Problem very efficiently with optimality guarantee. From the perspective of dynamical systems, a heuristic local search algorithm for an NP-complete combinatorial problem is a discrete dynamical system. In a local search system, an attractor drives the search trajectories into the vicinity of a globally optimal point in the solution space, and the convergence of local search trajectories makes the search system become a global and deterministic system. The attractor contains a small set of the most promising solutions to the problem. The attractor can reduce the problem size exponentially, and thus make the exhaustive search feasible. Therefore, this new search paradigm is called optimizing with attractor. The ABSS consists of two search phases: local search phase and exhaustive search phase. The local search process is used to quickly construct the attractor in the solution space, and the exhaustive search process is used to completely search the attractor to identify the optimal solution. Therefore, the exact optimal solution can be found quickly by combining local search and exhaustive search. This tutorial introduces the concept of an attractor in a local search system, and describes the process of optimizing with the attractor, using the Traveling Salesman Problem as the study platform.

Modern computing systems typically use a large number of independent, non-identical computing nodes to perform a set of coordinated computations in parallel. The computing system and its constituent computing nodes often exhibit more than two performance levels or states corresponding to different computing powers. This paper models and evaluates performability of large-scale multi-state computing systems, which is the probability that a computing system performs at a particular performance level. The heterogeneity in the constituent components of different nodes (due to factors such as different model generations, model suppliers, and operating environments) makes performability analysis difficult and challenging. In this paper a specification method for system performance level (SPL) is first introduced. A multi-valued decision diagram (MDD) based approach is then proposed for performability analysis of multi-state computing systems consisting of nodes with different state occupation probabilities, which encompasses novel and efficient MDD model generation procedures. Example and benchmark studies are performed to show that the proposed approach can offer efficient performability analysis of large-scale computing systems.

Performing a local search around solutions found by an evolutionary algorithm is a common practice. Local search is well known to significantly improve the solutions, in particular in the case of combinatorial problems. In this paper a new local search procedure is proposed that uses a locally established utopia point. In the tests in which several instances of the Travelling Salesman Problem (TSP) were solved using an evolutionary algorithm the proposed local search procedure outperformed a local search procedure based on Pareto dominance. Because the local search is focused on improving individual solutions and the multiobjective evolutionary algorithm can improve diversity, various strategies of sharing computational resources between the evolutionary algorithm and the local search are used in this paper. The results attained by the tested methods are compared with respect to computation time, which allows a fair comparison between strategies that distribute computational resources between the evolutionary optimization and the local search in various proportions.

Today, large scale computer systems have become an important component in production and scientific computing and lead to rapid advances in many disciplines. However, the size, and complexity of systems make them very difficult to detect unusual nodes automatically and traditional host monitoring tools are not capable of dealing with the need of anomaly detection in large amount of nodes. In this paper, we introduce a novel tool, LSAND, which could detect anomalistic nodes in a horizontal view of the machines with comparable configuration and tasks running on. We evaluated LSAND in a cluster environment, table the results of our experiment and give a discussion on the effect and show that LSAND is both effective and efficient for detecting anomalistic nodes with high-dimensional features.

<p>Unidata has developed and deployed data infrastructure and data-proximate scientific workflows and software tools using cloud computing technologies for accessing, analyzing, and visualizing geoscience data. These resources are provided to educators and researchers through the Unidata Science Gateway (https://science-gateway.unidata.ucar.edu) and deployed on the U. S. National Science Foundation funded Jetstream (https://jetstream-cloud.org) cloud facility. During the SARS-CoV-2/COVID-19 pandemic, the Unidata Science Gateway has been used by many universities to teach data-centric atmospheric science courses and conduct several software training workshops to advance skills in data science.<br><br>The COVID-19 pandemic led to the closure of university campuses with little advance notice. Educators at institutions of higher learning had to urgently transition from in-person teaching to online classrooms. While such a sudden change was disruptive for education, it also presented an opportunity to experiment with instructional technologies that have been emerging for the last few years. Web-based computational notebooks, with their mixture of explanatory text, equations, diagrams and interactive code are an effective tool for online learning. Their use is prevalent in many disciplines including the geosciences. Multi-user computational notebook servers (e.g., Jupyter Notebooks) enable specialists to deploy pre-configured scientific computing environments for the benefit of students. The use such tools and environments removes barriers for students who otherwise have to download and install complex software tools that can be time consuming to configure, simplifying workflows and reducing time to analysis and interpretation. It also provides a consistent computing environment for all students and democratizes access to resources. These servers can be provisioned with computational resources not found in a desktop computing setting and leverage cloud computing environments and high speed networks. They can be accessed from any web browser-enabled device like laptops and tablets.<br><br>Since spring 2020 when the Covid pandemic led to the closure of universities across the U. S., Unidata has assisted several earth science departments with computational notebook environments for their classes. We worked with educators to tailor these resources for their teaching objectives. We ensured the technology was correctly provisioned with appropriate computational resources and collaborated to have teaching material immediately available for students. There were many successful examples of online learning experiences.<br><br>In this paper, we describe the details of the Unidata Science Gateway resources and discuss how those resources enabled Unidata to support universities during the COVID-19 lockdown.</p>

Ant colony optimization (ACO) algorithm is a recent meta-heuristic method inspired by the behavior of real ant colonies. The algorithm uses parallel computation mechanism and performs strong robustness, but it faces the limitations of stagnation and premature convergence. In this paper, a hybrid PS-ACO algorithm, ACO algorithm modified by particle swarm optimization (PSO) algorithm, is presented. The pheromone updating rules of ACO are combined with the local and global search mechanisms of PSO. On one hand, the search space is expanded by the local exploration; on the other hand, the search process is directed by the global experience. The local and global search mechanisms are combined stochastically to balance the exploration and the exploitation, so that the search efficiency can be improved. The convergence analysis and parameters selection are given through simulations on traveling salesman problems (TSP). The results show that the hybrid PS-ACO algorithm has better convergence performance than genetic algorithm (GA), ACO and MMAS under the condition of limited evolution iterations.

A V-shape property of optimal schedule of jobs about a common due date

2D Bin packing problem (2DBPP) is an NP-hard combinatorial optimization problem. Multiobjective versions of this well-known industrial engineering problem can occur frequently in real world application. Recently, Hybrid Evolutionary Algorithms have appear as a new area of research with their ability to combine alternative heuristics and local search mechanisms together for higher quality solutions. In this study, we propose a set of novel multiobjective hybrid genetic and memetic algorithms that make use of the state-of-the-art metaheuristics and local search techniques for minimizing the number of bins while also maintaining the load balance. We analyze the optimization time and the resulting solution quality of the proposed algorithms on an offline 2DBPP benchmark problem set with 500 instances. Using these results of exhaustive experiments, we conclude that the proposed hybrid algorithms are robust with their ability to obtain a high percentage of the optimal solutions.

A well-planned arrangement of manufacturing departments on a two-dimensional planar region considerably increases the efficiency of its production systems, which is termed facility layout problem (FLP). Conventional layout design approach often designs intercell layout (determining exact location of each department on shop floor area) and flow path layout design of material handling system (MHS) step by step in a sequential manner. This results in suboptimal solutions for FLP. In this paper, an integrated approach is adopted to design the intercell layout and the flow path layout of MHS simultaneously. The quality of the final layout is evaluated by minimizing total material handling cost. Sequence pair (SP) representation is used for layout encoding. The translation from SP to layout is efficiently made by longest common subsequence (LCS) methodology. An elitist strategy genetic algorithm using simulated annealing (E-GASAA) as a local search mechanism is developed and tested with four test problem instances available in the literature. Elitist strategy is incorporated to enhance convergence characteristic of the proposed algorithm. It is found that the proposed E-GASAA is able to produce best solutions consistently for the test problem instance of different sizes within acceptable computational effort. In addition to that, we tried to reduce the computational load with the help of adopted LCS computation methodology and achieved a good improvement.

An Artificial Bee Colony Algorithm Based on Problem Data Properties for Scheduling Job Shops

In large-scale computer systems, deciding how to dispatch arriving jobs to servers is a primary factor affecting system performance. Consequently, there is a wealth of literature on designing, analyzing, and evaluating the performance of load balancing policies. For analytical tractability, most existing work on dispatching in large-scale systems makes a key assumption: that the servers are homogeneous, meaning that they all have the same speeds, capabilities, and available resources. But this assumption is not accurate in practice. Modern computer systems are instead heterogeneous: server farms may consist of multiple generations of hardware, servers with varied resources, or even virtual machines running in a cloud environment. Given the ubiquity of heterogeneity in today's systems, it is critically important to develop load balancing policies that perform well in heterogeneous environments. In this paper, we focus on systems in which server speeds are heterogeneous.