ChapterPDF Available

Abstract and Figures

Evolutionary algorithms (EAs) require large scale computing resources when tackling real world problems. Such computational requirement is derived from inherently complex fitness evaluation functions, large numbers of individuals per generation, and the number of iterations required by EAs to converge to a satisfactory solution. Therefore, any source of computing power can significantly benefit researchers using evolutionary algorithms. We present the use of volunteer computing (VC) as a platform for harnessing the computing resources of commodity machines that are nowadays present at homes, companies and institutions. Taking into account that currently desktop machines feature significant computing resources (dual cores, gigabytes of memory, gigabit network connections, etc.), VC has become a cost-effective platform for running time consuming evolutionary algorithms in order to solve complex problems, such as finding substructure in the Milky Way Galaxy, the problem we address in detail in this chapter.
Content may be subject to copyright.
A preview of the PDF is not available
... Following this line, during the last years there has been a growing interest in the use of EAs in distributed computing environments that move away from classical dedicated networks so common in the past. Among such environments we can cite cloud computing [19], P2P networks [14,28], or volunteer computing (VC) [7], just to name a few. ...
Conference Paper
Full-text available
We consider the deployment of island-based evolutionary algorithms (EAs) on irregular computational environments plagued with different kind of glitches. In particular we consider the effect that factors such as network latency and transient process suspensions have on the performance of the algorithm. To this end, we have conducted an extensive experimental study on a simulated environment in which the performance of the island-based EA can be analyzed and studied under controlled conditions for a wide range of scenarios in terms of both the intensity of glitches and the topology of the island-based model (scale-free networks and von Neumann grids are considered). It is shown that the EA is resilient enough to withstand moderately high latency rates and is not significantly affected by temporary island deactivations unless a fixed time-frame is considered. Combining both kind of glitches has a higher toll on performance, but the EA still shows resilience over a broad range of scenarios.
... In this sense, there has been during the last years an important focus on the use of EAs in emergent computational scenarios that depart from classical dedicated networks so common in the past. Among these we can cite cloud computing [13], P2P networks [21], or volunteer computing [5], just to name a few. The dynamic nature of the underlying computational substrate is one of the most distinguished features of some of these new scenarios -consider for example a P2P network in which nodes enter or leave the system subject to some uncontrollable dynamics caused by user interventions, network disruptions, eventual crashes, etc. ...
Chapter
Full-text available
We consider the deployment of island-based evolutionary algorithms (EAs) on unstable networks whose nodes exhibit correlated failures. We use the sandpile model in order to induce such complex, correlated failures in the system. A performance analysis is conducted, comparing the results obtained in both correlated and non-correlated scenarios for increasingly large volatility rates. It is observed that simple island-based EAs have a significant performance degradation in the correlated scenario with respect to its uncorrelated counterpart. However, the use of self-\(\star \) properties (self-scaling and self-sampling in this case) allows the EA to increase its resilience in this harder scenario, leading to a much more gentle degradation profile.
... Thus, the model was shown to perfectly work on desktop grids provided by the researchers. The surprise came when the model was applied using computing resources provided by volunteers under the well known volunteer computing model [14]. ...
Article
This chapter discusses the inherent parallel nature of evolutionary algorithms, and the role this parallelism can take when implementing them on different hardware architectures. We show the interest in studying ephemeral behaviors that distributed computing resources may feature and some EA’s self-properties of interest, such as the fault-tolerant nature that helps to fight the churn phenomenon. Moreover, interactive versions of EAs, which require distributed computing systems, allow to incorporate human based knowledge within the algorithm at different levels, providing new means for improving their computing capabilities while also requiring a proper analysis of human behavior under an EA framework. A proper understanding of ephemeral properties of hardware resources, human behavior in interactive applications and intrinsic parallel behaviors of population based algorithms will lead to significant improvements.
... Other researchers have exploited other network-based technologies to distribute an EA over a set of computing nodes. For instance, Cole et al. [8] uses the popular Berkeley Open Infrastructure for Network Computing (BOINC) to distribute an EA, using the volunteer computing model, where connected clients share idle CPU cycles with a research project. Another example is the work of Fernándezde-Vega et al. [15] , who also distributes multiple EA runs using a volunteer computing network through BOINC. ...
Article
Full-text available
This work presents the EvoSpace model for the development of pool-based evolutionary algorithms (Pool-EA). Conceptually, the EvoSpace model is built around a central repository or population store, incorporating some of the principles of the tuple-space model and adding additional features to tackle some of the issues associated with Pool-EAs; such as, work redundancy, starvation of the population pool, unreliability of connected clients or workers, and a large parameter space. The model is intended as a platform to develop search algorithms that take an opportunistic approach to computing, allowing the exploitation of freely available services over the Internet or volunteer computing resources within a local network. A comprehensive analysis of the model at both the conceptual and implementation levels is provided, evaluating performance based on efficiency, optima found and speedup, while providing a comparison with a standard EA and an island-based model. The issues of lost connections and system parametrization are studied and validated experimentally with encouraging results, that suggest how EvoSpace can be used to develop and implement different Pool-EAs for search and optimization.
... Modelado 3D de la Galaxia Milky Way a partir de datos del Sloan Digital Sky Survey [20]. ...
Conference Paper
Full-text available
En el presente trabajo hace un acercamiento a la Computación Distribuida, en particular, a la Computación de Alta Productividad. Se presenta el panorama actual de esta rama de la Ciencia de la Computación que está ganando espacio en el mundo de las herramientas computacionales que persiguen dar solución a problemas de alto costo computacional; se hace un recorrido que muestra cómo combinando el tiempo ocioso de nodos computacionales no dedicados es posible resolver problemas de este tipo; se abordan además los sistemas distribuidos más representativos de hoy en día. Finalmente se presentan los resultados obtenidos con la aplicación de uno de estos sistemas en el modelado de yacimientos lateríticos a partir de un modelo de Markov, problema de suma importancia para la Industria Cubana del Níquel.
Article
Solving complex real-world optimization problems is a computationally demanding task. To solve it efficiently and effectively, one must possess expert knowledge in various fields (problem domain knowledge, optimization, parallel and distributed computing) and appropriate expensive software and hardware resources. In this regard, we present a cloud-native, container-based distributed optimization framework that enables efficient and cost-effective optimization over platforms such as Amazon ECS/EKS, Azure AKS, and on-premise Kubernetes. The solution consists of dozens of microservices scaled out using a specially developed PETAS Auto-scaler based on predictive analytics. Existing schedulers, whether Kubernetes or commercial, do not take into account the specifics of optimization based on evolutionary algorithms. Therefore, their performance is not optimal in terms of results’ delivery time and cloud infrastructure costs. The proposed PETAS Auto-scaler elastically maintains an adequate number of worker pods following the exact pace dictated by the demands of the optimization process. We evaluate the proposed framework’s performance using two real-world computationally demanding optimizations. The first use case belongs to the manufacturing domain and involves optimization of the transportation pallets for train parts. The second use case belongs to the field of automated machine learning and includes neural architecture search and hyperparameter optimization. The results indicate an IaaS cost savings of up to 49% can be achieved, with almost unchanged result delivery time.
Article
Splitting a population into multiple instances is a technique used extensively in recent years to help improve the performance of nature-inspired optimization algorithms. Work on those populations can be done in parallel, and they can interact asynchronously, a fact that can be leveraged to create scalable implementations based on, among other methods, distributed, multi-threaded, parallel, and cloud-native computing. However, the design of these cloud-native, distributed, multi-population algorithms is not a trivial task. Using as a foundation monolithic (single-instance) solutions, adaptations at several levels, from the algorithmic to the functional, must be made to leverage the scalability, elasticity, (limited) fault-tolerance, reproducibility, and cost-effectiveness of cloud systems while, at the same time, conserving the intended functionality. Instead of an evolutive approach, in this paper, we propose a cloud-native optimization framework created from scratch, that can include multiple (population-based) algorithms without increasing the number of parameters that need tuning. This solution goes beyond the current state of the art, since it can support different algorithms at the same time, work asynchronously, and also be readily deployable to any cloud platform. We evaluate this solution’s performance and scalability, together with the effect other design parameters had on it, particularly the number and the size of populations with respect to problem size. The implemented platform is an excellent alternative for running locally or in the cloud, thus proving that cloud-native bioinspired algorithms perform better in their “natural” environment than other algorithms, and set a new baseline for scaling and performance of this kind of algorithms in the cloud.
Article
The cloud computing paradigm has gained wide acceptance in the scientific community, taking a significant share from fields previously reserved exclusively for High Performance Computing (HPC). On-demand access to a large amount of computing resources provided by Cloud makes it ideal for executing large-scale optimizations using evolutionary algorithms without the need for owning any computing infrastructure. In this regard, we extended WoBinGO, an existing parallel software framework for genetic algorithm based optimization, to be used in Cloud. With these extensions, the framework is capable of elastically and frugally utilizing the underlying cloud computing infrastructure for performing computationally expensive fitness evaluations. We studied two issues that are pertinent when dealing with large-scale optimization in the elastic cloud environment: the computing instance launching overhead and the price of engaging Cloud for solving optimization problems, in terms of the instances’ cumulative uptime. To explain the usability limits of WoBinGO framework running in the IaaS environment, a comprehensive analysis of the framework’s performance was given. Optimization of both total optimization time and total cumulative uptime, leads to minimizing the cost of cloud resources utilization. In this way, we are proposing an intelligent decision support engine based on artificial neural networks and metaheuristics to provide the user with an assessment of the framework’s behavior on the underlying infrastructure in terms of optimization duration and the cost of resource consumption. According to a given assessment, the user can decide upon faster delivery of results or lower infrastructure costs. The proposed software framework has been used to solve a complex real-world optimization problem of a subsurface rock mass model calibration. The results obtained from the private OpenStack deployment show that by using the proposed decision support engine, significant savings can be achieved in both optimization time and optimization cost.
Article
Research and development of automatic trading systems are becoming more frequent, as they can reach a high potential for predicting market movements. The use of these systems allows to manage a huge amount of data related to the factors that affect investment performance (macroeconomic variables, company information, industry indicators, market variables, etc.), while avoiding psychological reactions of traders when investing in financial markets.Movements in stock markets are continuous throughout each day, which requires trading systems must be supported by more powerful engines, since the amount of data to process grows, while the response time required to support operations is shortened. In this chapter we present two parallel implementations of a GA based trading system. The first uses a Grid Volunteer System based on BOINC and the second one takes advantage of a Graphic Processing Unit implementation.
Article
Genetic type particle methods are increasingly used to sample from complex high-dimensional distributions. They have found a wide range of applications in applied probability, Bayesian statistics, information theory, and engineering sciences. Understanding rigorously these new Monte Carlo simulation tools leads to fascinating mathematics related to Feynman-Kac path integral theory and their interacting particle interpretations. In this chapter, we provide an introduction to the stochastic modeling and the theoretical analysis of these particle algorithms. We also illustrate these methods through several applications.
Article
In this paper, we present the ParadisEO white-box object-oriented framework dedicated to the reusable design of parallel and distributed metaheuristics (PDM). ParadisEO provides a broad range of features including evolutionary algorithms (EA), local searches (LS), the most common parallel and distributed models and hybridization mechanisms, etc. This high content and utility encourages its use at European level. ParadisEO is based on a clear conceptual separation of the solution methods from the problems they are intended to solve. This separation confers to the user a maximum code and design reuse. Furthermore, the fine-grained nature of the classes provided by the framework allow a higher flexibility compared to other frameworks. ParadisEO is of the rare frameworks that provide the most common parallel and distributed models. Their implementation is portable on distributed-memory machines as well as on shared-memory multiprocessors, as it uses standard libraries such as MPI, PVM and PThreads. The models can be exploited in a transparent way, one has just to instantiate their associated provided classes. Their experimentation on the radio network design real-world application demonstrate their efficiency.
Article
Large scale surveys are providing vast amounts of data that can help us understand and study tidal debris more easily and accurately. A maximum likelihood method for determining the spatial properties of this tidal debris and the stellar Galactic spheroid has been developed to take advantage of these huge datasets. We present the results of studying the Sagittarius dwarf tidal stream in two SDSS stripes taken in the southern Galactic Cap using this method. This study was done using stars with the colors of blue F turnoff stars in SDSS. We detected Sagittarius debris at the positions (l,b,R) = (163.311°,-48.400°,30.23 kpc) and (l,b,R) = (34.775°,-72.342°,26.08 kpc). These debris pieces were found to have a FWHM of 6.53+/-0.54 kpc and 5.71+/-0.26 kpc and also to contain ~9,500 and ~16,700 F turnoff stars, respectively. The debris pieces were also found to have (spatial) directions of (Xˆ,Yˆ,Zˆ) = (0.758,0.254,-0.600) and (Xˆ,Yˆ,Zˆ) = (0.982,0.945,0.167), respectively. Using the results of the algorithm, we have also probabilistically separated the tidal debris from the stellar spheroid and present those results as well.
Article
Anyone with a home PC could join climate modellers in their attempt to forecast how the Earth's climate will evolve in the next century.