Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Evolvable Agents, a Fine Grained Approach for Distributed Evolutionary Computing: Walking towards the Peer-to-Peer Computing Frontiers

Juan Luis Jiménez Laredo, Pedro A. Castillo, Antonio M. Mora, Juan Julián Merelo

TL;DR

This paper introduces Evolvable Agents, a fine-grained, fully distributed evolutionary algorithm designed for heterogeneous networks and P2P environments. Each agent locally evolves a single solution while coordinating via a Black Board and a gossip-style Scheduler that self-adjusts migration rates according to link latency and bandwidth. Through three case studies against an Island-model baseline, the approach demonstrates improved time performance and robust scaling on small networks, with performance dependent on the benchmark function. The work lays groundwork for scalable dEAs in P2P settings and suggests directions for large-scale evaluation, dynamic population sizing, and integration with existing P2P frameworks like DREAM.

Abstract

In this work we propose a fine grained approach with self-adaptive migration rate for distributed evolutionary computation. Our target is to gain some insights on the effects caused by communication when the algorithm scales. To this end, we consider a set of basic topologies in order to avoid the overlapping of algorithmic effects between communication and topological structures. We analyse the approach viability by comparing how solution quality and algorithm speed change when the number of processors increases and compare it with an Island model based implementation. A finer-grained approach implies a better chance of achieving a larger scalable system; such a feature is crucial concerning large-scale parallel architectures such as Peer-to-Peer systems. In order to check scalability, we perform a threefold experimental evaluation of this model: First, we concentrate on the algorithmic results when the problem scales up to eight nodes in comparison with how it does following the Island model. Second, we analyse the computing time speedup of the approach while scaling. Finally, we analyse the network performance with the proposed self-adaptive migration rate policy that depends on the link latency and bandwidth. With this experimental setup, our approach shows better scalability than the Island model and a equivalent robustness on the average of the three test functions under study.

Evolvable Agents, a Fine Grained Approach for Distributed Evolutionary Computing: Walking towards the Peer-to-Peer Computing Frontiers

TL;DR

This paper introduces Evolvable Agents, a fine-grained, fully distributed evolutionary algorithm designed for heterogeneous networks and P2P environments. Each agent locally evolves a single solution while coordinating via a Black Board and a gossip-style Scheduler that self-adjusts migration rates according to link latency and bandwidth. Through three case studies against an Island-model baseline, the approach demonstrates improved time performance and robust scaling on small networks, with performance dependent on the benchmark function. The work lays groundwork for scalable dEAs in P2P settings and suggests directions for large-scale evaluation, dynamic population sizing, and integration with existing P2P frameworks like DREAM.

Abstract

In this work we propose a fine grained approach with self-adaptive migration rate for distributed evolutionary computation. Our target is to gain some insights on the effects caused by communication when the algorithm scales. To this end, we consider a set of basic topologies in order to avoid the overlapping of algorithmic effects between communication and topological structures. We analyse the approach viability by comparing how solution quality and algorithm speed change when the number of processors increases and compare it with an Island model based implementation. A finer-grained approach implies a better chance of achieving a larger scalable system; such a feature is crucial concerning large-scale parallel architectures such as Peer-to-Peer systems. In order to check scalability, we perform a threefold experimental evaluation of this model: First, we concentrate on the algorithmic results when the problem scales up to eight nodes in comparison with how it does following the Island model. Second, we analyse the computing time speedup of the approach while scaling. Finally, we analyse the network performance with the proposed self-adaptive migration rate policy that depends on the link latency and bandwidth. With this experimental setup, our approach shows better scalability than the Island model and a equivalent robustness on the average of the three test functions under study.
Paper Structure (21 sections, 6 equations, 15 figures, 1 table, 4 algorithms)

This paper contains 21 sections, 6 equations, 15 figures, 1 table, 4 algorithms.

Table of Contents

  1. Introduction
  2. State of the art
  3. Overall Model Description
  4. Blackboard
  5. Scheduler Agent and Cache
  6. Evolvable Agent with Tournament Selection
  7. Methods and Experiments
  8. The Benchmark
  9. Shifted Sphere function.
  10. Shifted Rotated Rastrigin's function.
  11. Schwefel function.
  12. Experimental Setup
  13. Experimental Results
  14. Case Study 1: Evolvable Agents vs. Island model
  15. Shifted Sphere function.
  16. ...and 6 more sections

Figures (15)

  • Figure 1: Overall architecture of the model
  • Figure 2: Format of a cache entry. It provides the following information about a foreign node: Address, number of evaluations performed and one individual of its population termed solution
  • Figure 3: Sphere function in two dimensions (Suganthan et al., 2005)
  • Figure 4: Rastrigin function (Suganthan et al., 2005)
  • Figure 5: Schwefel function (Suganthan et al., 2005)
  • ...and 10 more figures