Implementing Parallel Differential Evolution on Spark

  title={Implementing Parallel Differential Evolution on Spark},
  author={Diego Teijeiro and Xo{\'a}n C. Pardo and Patricia Gonz{\'a}lez and Julio R. Banga and Ram{\'o}n Doallo},
Metaheuristics are gaining increased attention as an efficient way of solving hard global optimization problems. Differential Evolution (DE) is one of the most popular algorithms in that class. However, its application to realistic problems results in excessive computation times. Therefore, several parallel DE schemes have been proposed, most of them focused on traditional parallel programming interfaces and infrastructures. However, with the emergence of Cloud Computing, new programming models… 

Evaluation of Parallel Differential Evolution Implementations on MapReduce and Spark

This paper investigates how parallel metaheuristics deal with new programming models to deal with large scale computations on commodity clusters and Cloud resources by the parallelization of the popular Differential Evolution algorithm using MapReduce and Spark.

Towards cloud-based parallel metaheuristics

This paper has developed, using Spark, an island-based parallel version of Differential Evolution, a simple population-based metaheuristic that, at the same time, is very popular for being very efficient in real function global optimization.

A cloud-based enhanced differential evolution algorithm for parameter estimation problems in computational systems biology

This paper proposes a parallel implementation of an enhanced DE using Spark, which drastically reduces the execution time, by means of including a selected local search and exploiting the available distributed resources.

A Spark-based differential evolution with grouping topology model for large-scale global optimization

The results demonstrate that the SgtDE, especially a combination of better topology and migration strategy, exhibits significant performance in applying various DE variants, and can act as the next generation optimizer of the cloud computing platform.

A Hybrid Mechanism of Particle Swarm Optimization and Differential Evolution Algorithms based on Spark

This paper proposes a hybrid mechanism of particle swarm optimization (PSO) and differential evolution (DE) algorithms based on Spark (SparkPSODE) and shows that, in comparison with several algorithms, the proposed SparkPSODE algorithm obtains better optimization performance through experimental results.

Multimethod optimization in the cloud: A case‐study in systems biology modelling

This paper proposes a self‐adaptive cooperative parallel multimethod for global optimization, which aims to perform a thorough exploration of the solution space by means of multiple concurrent executions of a broad range of search strategies.

Hybrid parallel multimethod hyperheuristic for mixed-integer dynamic optimization problems in computational systems biology

A hybrid parallel scheme of the multimethod, using both message-passing (MPI) and shared memory (OpenMP) models, is presented and evaluated, achieving significant performance improvements versus the sequential and the non-cooperative parallel solutions.

A Distributed Quantum-Behaved Particle Swarm Optimization Using Opposition-Based Learning on Spark for Large-Scale Optimization Problem

A distributed quantum-behaved particle swarm optimization algorithm (SDQPSO) based on the SDCEA framework has high scalability, which can be applied to solve the large-scale optimization problem.

Evolutionary Induction of Classification Trees on Spark

This paper investigates the application of Spark to speed up an evolutionary induction of classification trees in the Global Decision Tree (GDT) system and develops a Java-based module for Spark-based acceleration of the most computationally demanding fitness evaluation.



Parallel metaheuristics: recent advances and new trends

The state of the art in parallel metaheuristics is discussed here on, in a summarized manner, to provide a solution to deal with some of the growing topics.

Concurrent Differential Evolution Based on MapReduce

Through the numerical experiment conducted on a wide range of benchmark problems, the speedup of DE due to the use of multiple cores is demonstrated and the goodness of the proposed concurrent implementation of DE is examined and proved with respect to four categories, namely efficiency, simplicity, portability and scalability.

Designing a parallel evolutionary algorithm for inferring gene networks on the cloud computing environment

Parallel population-based algorithms can effectively determine network parameters and they perform better than the widely-used sequential algorithms in gene network inference and can be distributed to the cloud computing environment to speed up the computation.

Fast parallelization of differential evolution algorithm using MapReduce

This paper demonstrates how to apply MapReduce and the open source Hadoop framework for a quick and easy parallelization of the Differential Evolution algorithm.

Distributed Simulated Annealing with MapReduce

This paper designs six algorithmic patterns of distributed simulated annealing with MapReduce, instantiate the patterns into MR implementations to solve a sample TSP problem, and evaluates the solution quality and the speedup of the implementations on a cloud computing platform, Amazon's Elastic Map reduce.

Parallel Metaheuristics: A New Class of Algorithms

This chapter discusses Metaheuristics and Parallelism in Telecommunications, which has applications in Bioinformatics and Parallel Meta heuristics, and Theory of Parallel Genetic Algorithms, which focuses on the latter.

Scaling Genetic Algorithms Using MapReduce

This paper describes the algorithm design and implementation of GAs on Hadoop, an open source implementation of MapReduce, and demonstrates the convergence and scalability up to 10^5 variable problems.

Parallel PSO using MapReduce

This work describes MapReduce and shows how PSO can be naturally expressed in this model, without explicitly addressing any of the details of parallelization, and demonstrates that MRPSO scales to 256 processors on moderately difficult problems and tolerates node failures.

Spinning Fast Iterative Data Flows

This work proposes a method to integrate incremental iterations, a form of workset iterations, with parallel dataflows and presents an extension to the programming model for incremental iterations that alleviates for the lack of mutable state in dataflow and allows for exploiting the sparse computational dependencies inherent in many iterative algorithms.