Parallel Evolutionary Algorithm Research Articles

Development of Parallel Algorithms for Probabilistic Models Learning for Web Application Testing

In the current conditions of testing methods and algorithms development to combine individual test components in the form of a hierarchical model reflecting the connections and states between these components, as well as allowing to evaluate the probability of a transition between model states in the event of information about the successful implementation of a certain test to detect a program error is of particular importance. Existing approaches do not allow optimal adjustment of model parameters, as well as qualitative calculation and establishment of directions of connections between individual components of this model. The scientific study considered the possibility of using effective numerical methods to solve the problem of training probabilistic models built on the basis of Bayesian networks to solve the main problems of testing web applications. The main approaches to create parallel algorithms implementing the main functional capabilities underlying the implementation of the procedure for training test models are considered. Analysis of the developed training algorithms effectiveness for testing certain groups of program errors allowing the most optimal parameters of the DBS model was made, as well as a justification for their use in distributed data processing systems. We have developed algorithmic solutions for Jacobi and Hesse matrices calculation optimization based on Cannon and Fox algorithms. The process of testing errors of web application is simulated and presented in the form of a dynamic Bayesian network obtained from the results of the structure and parameters learning procedure. The validity of all theoretical results is confirmed by a large number of experimental results proving the validity of the put forward assumptions, testing methods and models presented in the form of dynamic Bayesian networks.

Scaling and Benchmarking an Evolutionary Algorithm for Constructing Biophysical Neuronal Models

Single neuron models are fundamental for computational modeling of the brain's neuronal networks, and understanding how ion channel dynamics mediate neural function. A challenge in defining such models is determining biophysically realistic channel distributions. Here, we present an efficient, highly parallel evolutionary algorithm for developing such models, named NeuroGPU-EA. NeuroGPU-EA uses CPUs and GPUs concurrently to simulate and evaluate neuron membrane potentials with respect to multiple stimuli. We demonstrate a logarithmic cost for scaling the stimuli used in the fitting procedure. NeuroGPU-EA outperforms the typically used CPU based evolutionary algorithm by a factor of 10 on a series of scaling benchmarks. We report observed performance bottlenecks and propose mitigation strategies. Finally, we also discuss the potential of this method for efficient simulation and evaluation of electrophysiological waveforms.

Parallel Genetic Algorithm Based on Construction of Gene Pool in the Ordinary Network for TSP

Though using parallel evolutionary algorithm to solve large-scale TSP problems is efficient, the parallel computer costs too much and the algorithm is not easy to expand. To address this issue, I propose a parallel genetic algorithm based on a gene pool under the existing network. To replace the group-genes in the evolutionary algorithm with the genes from the gene pool, the algorithm conducts greedy algorithm. The host process conducts greedy algorithm and improved evolutionary algorithm of Inver-over operator while the child process performs the improved hybrid genetic algorithms. Simulation results demonstrate that this algorithm achieves a better solution.

On the numerical approximation of some inverse problems governed by nonlinear delay differential equation

The paper deals with the approximate solving of an inverse problem for the nonlinear delay differential equation, which consists of finding the initial moment and delay parameter based on some observed data. The inverse problem is considered as a nonlinear optimal control problem for which the necessary conditions of optimality are formulated and proved. The obtained optimal control problem is solved by a method based on an improved parallel evolutionary algorithm. The efficiency of the proposed approach is demonstrated through various numerical experiments.

Development of Parallel Algorithms for Intelligent Transportation Systems

This paper deals with the creation of parallel algorithms implementing macro-and microscopic traffic flow models on modern supercomputers. High-performance computing contributes to the development of intelligent transportation systems based on information technologies and aimed at the effective regulation of traffic in large cities. As a macroscopic approach, the quasi-gas-dynamic traffic model approximated by explicit finite-difference schemes is proposed. One- and two-dimensional variants of the system are considered, and the concept of lateral velocity and different equations for obtaining it are discussed. The microscopic approach is represented by the multilane cellular automata model. The previously developed model is extended to reproduce synchronized flow in accordance with Kerner’s three-phase theory. The new version starts from the Kerner–Klenov–Schreckenberg–Wolf model and operates with the concept of the synchronization gap. Macroscopic models are relevant for determining the common characteristics of road traffic, while microscopic models are useful for a detailed description of cars’ movement. Both approaches possess inner parallelism. The parallel algorithms are based on the geometrical parallelism principle with different boundary conditions at interfaces of the subdomains. Sufficiently high speedups were reached when up to 100 processors were involved in calculations. The proposed algorithms can serve as the core of ITS.

An optimization approach to the development of parallel algorithms for digital signal processing operations

The paper studies the optimization possibilities of an approach which made it possible to develop a formal tool, i.e. a model of computation organization for digital signal processing (DSP) operations and operations structurally similar to them. The resulting formal tool describes the inter-nal parameterized structure of operations and generates adaptive algorithms that can adjust to dif-ferent conditions of parallel computation. The developed approach to the construction of such algo-rithms endows them with extended functionality, which ensures the implementation of the follow-ing capabilities: changes in the parameters of algorithm structures; synthesis of their variety; opti-mization of variants of computation organization. Due to the variety of this functionality, several directions are identified for implementing this optimization. A general description of the optimiza-tion approach to the reasonable choice of the best variant of computation organization under the given conditions of their implementation is given. We consider a scheme of actions aimed at com-putation optimization and enabling you to develop various classes of parameterized parallel algo-rithms within the framework of the approach proposed.

Transfer Learning-Based Parallel Evolutionary Algorithm Framework for Bilevel Optimization

Evolutionary algorithms (EAs) have been recognized as a promising approach for bilevel optimization. However, the population-based characteristic of EAs largely influences their efficiency and effectiveness due to the nested structure of the two levels of optimization problems. In this article, we propose a transfer learning-based parallel EA (TLEA) framework for bilevel optimization. In this framework, the task of optimizing a set of lower level problems parameterized by upper level variables is conducted in a parallel manner. In the meanwhile, a transfer learning strategy is developed to improve the effectiveness of each lower level search (LLS) process. In practice, we implement two versions of the TLEA: the first version uses the covariance matrix adaptation evolutionary strategy and the second version uses the differential evolution as the evolutionary operator in lower level optimization. The experimental studies on two sets of widely used bilevel optimization benchmark problems are conducted, and the performance of the two TLEA implementations is compared to that of four well-established evolutionary bilevel optimization algorithms to verify the effectiveness and efficiency of the proposed algorithm framework.

Parallel algorithm development and testing using Petri-object simulation

Parallel algorithms are problematic to develop because of the negative influence of synchronisation, complicated behaviour of threads’ capturing computing resources. Experimental results show performance time’s strong dependence on algorithm parameters, such as the number of subtasks and the complexity of each task. The optimal value of subtask complexity is revealed for the particular algorithm. It is the same for different complexity of the parallelised task (with the same computing resource). To guarantee algorithm speed-up it is important to have a method for investigating the efficiency of parallel algorithm before its implementation on specified computing resources. Stochastic Petri net potentially could be a high accuracy tool for investigating the efficiency of a parallel algorithm. However, a huge number of elements are needed to compose a model of non-trivial algorithm that limits the application of this tool in practice. Petri-object simulation method allows replication of Petri nets with specified parameters and model creation of a list of linked Petri-objects. Basic templates for the model creation of a multithreaded algorithm are developed. Applying these templates, the model of the parallel discrete event simulation algorithm is developed and investigated. By the model results, the algorithm parameters providing the least performance time can be determined.

Mixed Integer Programming versus Evolutionary Computation for Optimizing a Hard Real-World Staff Assignment Problem

Assigning staff to engagements according to hard constraints while optimizing several objectives is a task encountered by many companies on a regular basis. Simplified versions of such assignment problems are NP-hard. Despite this, a typical approach to solving them consists of formulating them as mixed integer programming (MIP) problems and using a stateof-the-art solver to get solutions that closely approximate the optimum.In this paper, we consider a complex real-world staff assignment problem encountered by the professional service company KPMG, with the goal of finding an algorithm that solves it faster and with a better solution than a commercial MIP solver. We follow the evolutionary algorithm (EA) metaheuristic and design a search heuristic which iteratively improves a solution using domain-specific mutation operators. Furthermore, we use a flow algorithm to optimally solve a subproblem, which tremendously reduces the search space for the EA.For our real-world instance of the assignment problem, given the same total time budget of 100 hours, a parallel EA approach finds a solution that is only 1.7% away from an upper bound for the (unknown) optimum within under five hours, while the MIP solver Gurobi still has a gap of 10.5%.

Clone Chaotic Parallel Evolutionary Algorithm for Low-Energy Clustering in High-Density Wireless Sensor Networks

Because the sensors are constrained in energy capabilities, low-energy clustering has become a challenging problem in high-density wireless sensor networks (HDWSNs). Usually, sensor nodes tend to be tiny devices along with constrained clustering abilities. To have a low communication energy consumption, a low-energy clustering scheme should be designed properly. In this work, a new cloned chaotic parallel evolution algorithm (CCPEA) is proposed, and a low-energy clustering model is established to lower the communication energy consumption of HDWSNs. By introducing CCPEA into the low-energy clustering, an objective function is designed for evaluating the communication energy consumption. For this problem, we define a clone operator to minimize the communication energy consumption of HDWSNs, use the chaotic operator to randomly generate the initial population to expand the search range to avoid local optimization, and find the parallel operator to speed up the convergence speed. In the experiment, the effect of CCPEA is compared to heuristic approaches of particle swarm optimization (PSO) and simulated annealing (SA) for the HDWSNs with different numbers of sensors. Simulation experiments demonstrate that the presented CCPEA method achieves a lower communication energy consumption and faster convergence speed than PSO and SA.

On Development of Parallel Algorithms for Solving Parabolic and Elliptic Equations

In this paper, we present results of the development of certain parallel numerical methods for solving three-dimensional evolutionary and stationary problems of diffusion and heat transfer. We present a detailed description of a special, explicit iteration scheme for parabolic equations and discuss a multigrid technology used for solving elliptic equations and implicit schemes for parabolic equations.

Guided parallelized stochastic gradient descent for delay compensation

Stochastic gradient descent (SGD) algorithm and its variations have been effectively used to optimize neural network models. However, with the rapid growth of big data and deep learning, SGD is no longer the most suitable choice due to its natural behavior of sequential optimization of the error function. This has led to the development of parallel SGD algorithms, such as asynchronous SGD (ASGD) and synchronous SGD (SSGD) to train deep neural networks. However, it introduces a high variance due to the delay in parameter (weight) update. We address this delay in our proposed algorithm and try to minimize its impact. We employed guided SGD (gSGD) that encourages consistent examples to steer the convergence by compensating the unpredictable deviation caused by the delay. Its convergence rate is also similar to A/SSGD, however, some additional (parallel) processing is required to compensate for the delay. The experimental results demonstrate that our proposed approach has been able to mitigate the impact of delay for the quality of classification accuracy. The guided approach with SSGD clearly outperforms sequential SGD and even achieves an accuracy close to sequential SGD for some benchmark datasets.

Mean-CVaR Portfolio Optimization Approaches with Variable Cardinality Constraint and Rebalancing Process

This work compares Mean-CVaR portfolio optimization models with variable cardinality constraint and rebalancing process. It considers integer and continuous decision variables, the number of asset lots and asset investment rate, respectively, and the linear and non-linear formulations of CVaR. Exact methods are used to solve the linear models and parallel evolutionary algorithms are used to solve the non-linear models. The in-sample analysis compares the sets of multiobjective optimization solutions, evaluating the effect of the cardinality of portfolios, with respect to the returns and risks. The out-of-sample analysis performs simulations with stock market trading, considering historical data with different data granularity and transaction costs, aiming to analyze the effects of these characteristics on the financial risks and gains. Results show that models considering asset lots are more effective in practice and that the exact methods provide solutions closer to the heuristics, with greater execution time. Out-of-sample analysis indicates the robustness of the portfolio optimization models pointing out similar behavior of financial gains for different values of transaction costs. Optimization with higher granularity provides greater risk, but also offers chances of high profits.

Parallel Differential Evolutionary Particle Filtering Algorithm Based on the CUDA Unfolding Cycle

Aiming at the problem of low statute efficiency of prefix sum execution during the execution of the parallel differential evolutionary particle filtering algorithm, a filtering algorithm based on the CUDA unfolding cyclic prefix sum is proposed to remove the thread differentiation and thread idleness existing in the parallel prefix sum by unfolding the cyclic method and unfolding the thread bundle method, optimize the cycle, and improve the prefix sum execution efficiency. By introducing the parallel strategy, the differential evolutionary particle filtering algorithm is implemented in parallel and executed on the GPU side using the improved prefix sum computation during the algorithm update. Through big data analysis, the results show that this parallel differential evolutionary particle filtering algorithm with the improved prefix sum statute can effectively improve differential evolutionary particle filtering for nonlinear system states and real‐time performance in heterogeneous parallel processing systems.

Improving CPU utilization of interleaving generation parallel evolutionary algorithm with precedence evaluation of tentative solutions and their suspension

This paper proposes a new mechanism to improve the CPU efficiency of parallel evolutionary algorithms (PEAs). The proposed method is based on interleaving generation evolutionary algorithm (IGEA) t...

Development of Time-Division Parallel Algorithm and Its Application to Designing Electronic Devices

Development of Time-Division Parallel Algorithm and Its Application to Designing Electronic Devices

A parallel evolutionary multiple-try metropolis Markov chain Monte Carlo algorithm for sampling spatial partitions

We develop an Evolutionary Markov Chain Monte Carlo (EMCMC) algorithm for sampling spatial partitions that lie within a large and complex spatial state space. Our algorithm combines the advantages of evolutionary algorithms (EAs) as optimization heuristics for state space traversal and the theoretical convergence properties of Markov Chain Monte Carlo algorithms for sampling from unknown distributions. Local optimality information that is identified via a directed search by our optimization heuristic is used to adaptively update a Markov chain in a promising direction within the framework of a Multiple-Try Metropolis Markov Chain model that incorporates a generalized Metropolis-Hasting ratio. We further expand the reach of our EMCMC algorithm by harnessing the computational power afforded by massively parallel architecture through the integration of a parallel EA framework that guides Markov chains running in parallel.

The approach to increasing the efficiency of convolution algorithms in modern high-availability systems

Most modern high-availability information systems are either based on systems that are classified as artificial intelligence, or to a large extent include them as components. The solution of many problems of artificial intelligence is based on the use of algorithms that implement the convolution operation (for example, algorithms for training neural networks). The article proposes an approach that, on the one hand, provides a strict formalization of this operation based on the algebra of multidimensional matrices, and on the other hand, due to this, it provides such practical advantages as simplifying and reducing the cost of developing such information systems, as well as reducing the execution time of queries to them. due to the simplicity of the development of parallel algorithms and programs and the efficient use of parallel computing systems. Convolution is an indispensable operation for solving many scientific and technical problems, such as machine learning, data analysis, signal processing, image processing filters. An important role is played by multidimensional convolutions, which are widely used in various subject areas. At the same time, due to the complexity of the algorithms that implement them, in practice even three-dimensional convolutions are used much less frequently than one- and two-dimensional ones. Replacing a multidimensional convolution operation with a sequence of convolution operations of lower dimensions significantly increases its computational complexity. The main reason for this lies in the absence of a unified strict definition of the operation and overload in mathematics of the term «convolution». Therefore, the article discusses a multidimensional-matrix computation model, which allows one to effectively formalize problems whose solution uses multidimensional convolution operations, and to implement an effective solution to these problems due to the natural parallelism inherent in the operations of the algebra of multidimensional matrices.

Diversified Personalized Recommendation Optimization Based on Mobile Data

With the advent of the Internet of Things, especially the Internet of Vehicles, abundant environmental and mobile data can be generated continuously. A personalized recommender system is one of the important methods for solving the problem of big data overload. However, to make use of these mobile data from vehicles, traditional recommender services are confronted by severe challenges. Therefore, we study the diversified recommendation problem based on a real-world dataset, represented as a tensor with three dimensions of user, location and activity. As the tensor is rather sparse, we employ tensor decomposition to predict missing values. Additionally, we directly regard recommendation precision as an objective. In addition to precision, we also consider the recommendation novelty and coverage, providing a more comprehensive view of the recommender system. Thus, visitors can discover attractive spots that are less visited in a personalized manner, relieving traffic pressure at famous scenic spots and balancing overall transportation. By integrating all these objectives, we construct a many-objective recommendation model. To optimize this model, we propose a distributed parallel evolutionary algorithm employing the nondominated ranking and crowding distance. Compared with the state-of-the-art algorithms, the proposed algorithm performs well and is very efficient.

Synchronous and asynchronous parallel computation for large-scale optimal control of connected vehicles

Connected vehicles is an important intelligent transportation system to improve the traffic performance. This paper proposes two parallel computation algorithms to solve a large-scale optimal control problem in the coordination of multiple connected vehicles. The coordination is formulated as a centralized optimization problem in the receding horizon fashion. A decentralized computation network is designed to facilitate the development of parallel algorithms. We use Taylor series to linearize non-convex constraints, and introduce a set of consensus constraints to transform the centralized problem to a standard consensus optimization problem. A synchronous parallel algorithm is firstly proposed to solve the consensus optimization problem by applying the alternating direction method of multipliers (ADMM). The ADMM framework allows us to decompose the coupling constraints and decision variables, leading to parallel iterations for each vehicle in a synchronous fashion. We then propose an asynchronous version of the parallel algorithm that allows the vehicles to update their variables asynchronously in the computation network. The effectiveness and efficiency of the proposed algorithms are validated by extensive numerical simulations.