Designing robust volunteer-based evolutionary algorithms

Abstract

This paper tackles the design of scalable and fault-tolerant evolutionary algorithms computed on volunteer platforms. These platforms aggregate computational resources from contributors all around the world. Given that resources may join the system only for a limited period of time, the challenge of a volunteer-based evolutionary algorithm is to take advantage of a large amount of computational power that in turn is volatile. The paper analyzes first the speed of convergence of massively parallel evolutionary algorithms. Then, it provides some guidance about how to design efficient policies to overcome the algorithmic loss of quality when the system undergoes high rates of transient failures, i.e. computers fail only for a limited period of time and then become available again. In order to provide empirical evidence, experiments were conducted for two well-known problems which require large population sizes to be solved, the first based on a genetic algorithm and the second on genetic programming. Results show that, in general, evolutionary algorithms undergo a graceful degradation under the stress of losing computing nodes. Additionally, new available nodes can also contribute to improving the search process. Despite losing up to 90% of the initial computing resources, volunteer-based evolutionary algorithms can find the same solutions in a failure-prone as in a failure-free run.

WebAL Comes of Age: A Review of the First 21 Years of Artificial Life on the Web

“…Svarga, created by Second Life user Laukosargas Svarog, was an island with a fully functioning ecosystem comprising a weather system and various types of plants and animals. 49 The island can still be visited in Second Life today, 50 but it was purchased from Svarog by Linden Lab in 2010 51 and some of its original ecosystem features may no longer be present. Shortly after the release of Svarga, a separate effort was launched by the Ecosystem Working Group and associated with the in-game location Terminus.…”

“…These include studies employing popular cloud-based storage services (specifically, Dropbox 136 and SugarSync 137 ) [67], and others implementing JavaScript-based evolutionary algorithms [66,65]. Other work has studied how the performance of different architectures of web-based evolutionary algorithms is affected when client nodes join and leave the cluster during an evolutionary run [51]. These studies present some interesting ideas that could be applied to WebAL projects.…”

The human in the loop: volunteer-based metacomputers as a socio-technical system

“…The number of users is key in the performance of these systems, but it also essential to adapt the algorithm itself to the available resources, as shown by (Milani, 2004), although EAs can be readily distributed via population splitting or by farming out the evaluation to all the nodes available. However, user churn affects experiment performance (González Lombraña et al, 2010;Nogueras and Cotta, 2015) and also the performance of the algorithm itself (Laredo et al, 2014). All these issues imply that a the performance of a volunteer system cannot be measured without first understanding its dynamics.…”

“…Some of the essential metrics in volunteer computing like the number of users or the time spent by every one in the computation in browser-based volunteer computing experiments, have only been studied in a limited way in (Laredo et al, 2014) on the basis of a single run. Studies using volunteer computing platforms such as SETI@home (Javadi et al, 2009;Merelo et al, 2008) found out that the Weibull, log-normal and Gamma distribution modeled quite well the availability of resources in several clusters of that framework; the shape of those distributions is a skewed bell with more resources in the low areas than in the high areas: there are many users that give a small amount of cycles, while there are just a few that give many cycles.…”

Resilience within Ultrascale Computing System: Challenges and Opportunities from Nesus Project

“…A set of recent studies [20,26,30,35,42,43] illustrate what seems to be a natural resilience of EAs against a model of destructive failures (crash failures). With a properly designed execution, the system experiences a graceful degradation [35]. This means, that up to some threshold and despite the failures, the results are still delivered.…”