Parallel Simulated Annealing Algorithm Research Articles

Investigation of temperature parallel simulated annealing for optimizing continuous functions with application to hyperspectral tomography

The simulated annealing (SA) algorithm is a well-established optimization technique which has found applications in many research areas. However, the SA algorithm is limited in its application due to the high computational cost and the difficulties in determining the annealing schedule. This paper demonstrates that the temperature parallel simulated annealing (TPSA) algorithm, a parallel implementation of the SA algorithm, shows great promise to overcome these limitations when applied to continuous functions. The TPSA algorithm greatly reduces the computational time due to its parallel nature, and avoids the determination of the annealing schedule by fixing the temperatures during the annealing process. The main contributions of this paper are threefold. First, this paper explains a simple and effective way to determine the temperatures by applying the concept of critical temperature (TC). Second, this paper presents systematic tests of the TPSA algorithm on various continuous functions, demonstrating comparable performance as well-established sequential SA algorithms. Third, this paper demonstrates the application of the TPSA algorithm on a difficult practical inverse problem, namely the hyperspectral tomography problem. The results and conclusions presented in this work provide are expected to be useful for the further development and expanded applications of the TPSA algorithm.

Load balancing on temporally heterogeneous cluster of workstations for parallel simulated annealing

Simulated annealing (SA) is a general-purpose optimization technique widely used in various combinatorial optimization problems. However, the main drawback of this technique is a long computation time required to obtain a good quality of solution. Clusters have emerged as a feasible and popular platform for parallel computing in many applications. Computing nodes on many of the clusters available today are temporally heterogeneous. In this study, multiple Markov chain (MMC) parallel simulated annealing (PSA) algorithms have been implemented on a temporally heterogeneous cluster of workstations to solve the graph partitioning problem and their performance has been analyzed in detail. Temporal heterogeneity of a cluster of workstations is harnessed by employing static and dynamic load balancing techniques to further improve efficiency and scalability of the MMC PSA algorithms.

BWR in-core fuel management optimization using parallel simulated annealing in FORMOSA-B

The process of finding optimized fuel reload patterns for boiling water reactors is complicated by a number of factors including the large number of fuel assemblies involved, the three-dimensional neutronic and thermal-hydraulic variations, and the interplay of coolant flow rate with control rod programming. The FORMOSA-B code was developed to provide an automated method for finding fuel loading patterns, control rod programs and coolant flow rate schedules to minimize certain quantitative metrics of core performance while satisfying given operational constraints. One drawback of this code has been the long runtimes required for a complete cycle optimization on a desktop workstation (oftentimes several days or more). To address this shortcoming, a parallel simulated annealing algorithm has been added to the FORMOSA-B code, so that the runtimes may be greatly reduced by using a multiprocessor computer cluster. Tests of the algorithm on a sample problem indicate that it is capable of parallel efficiencies exceeding 80% when using four processors.

Reverse engineering a gene network using an asynchronous parallel evolution strategy

BackgroundThe use of reverse engineering methods to infer gene regulatory networks by fitting mathematical models to gene expression data is becoming increasingly popular and successful. However, increasing model complexity means that more powerful global optimisation techniques are required for model fitting. The parallel Lam Simulated Annealing (pLSA) algorithm has been used in such approaches, but recent research has shown that island Evolutionary Strategies can produce faster, more reliable results. However, no parallel island Evolutionary Strategy (piES) has yet been demonstrated to be effective for this task.ResultsHere, we present synchronous and asynchronous versions of the piES algorithm, and apply them to a real reverse engineering problem: inferring parameters in the gap gene network. We find that the asynchronous piES exhibits very little communication overhead, and shows significant speed-up for up to 50 nodes: the piES running on 50 nodes is nearly 10 times faster than the best serial algorithm. We compare the asynchronous piES to pLSA on the same test problem, measuring the time required to reach particular levels of residual error, and show that it shows much faster convergence than pLSA across all optimisation conditions tested.ConclusionsOur results demonstrate that the piES is consistently faster and more reliable than the pLSA algorithm on this problem, and scales better with increasing numbers of nodes. In addition, the piES is especially well suited to further improvements and adaptations: Firstly, the algorithm's fast initial descent speed and high reliability make it a good candidate for being used as part of a global/local search hybrid algorithm. Secondly, it has the potential to be used as part of a hierarchical evolutionary algorithm, which takes advantage of modern multi-core computing architectures.

Three-dimensional reconstruction of statistically optimal unit cells of polydisperse particulate composites from microtomography

In this paper, we present a systematic approach for characterization and reconstruction of statistically optimal representative unit cells of polydisperse particulate composites. Microtomography is used to gather rich three-dimensional data of a packed glass bead system. First-, second-, and third-order probability functions are used to characterize the morphology of the material, and the parallel augmented simulated annealing algorithm is employed for reconstruction of the statistically equivalent medium. Both the fully resolved probability spectrum and the geometrically exact particle shapes are considered in this study, rendering the optimization problem multidimensional with a highly complex objective function. A ten-phase particulate composite composed of packed glass beads in a cylindrical specimen is investigated, and a unit cell is reconstructed on massively parallel computers. Further, rigorous error analysis of the statistical descriptors (probability functions) is presented and a detailed comparison between statistics of the voxel-derived pack and the representative cell is made.

Investigation of temperature parallel simulated annealing for optimizing continuous functions

The simulated annealing (SA) algorithm is a well-established optimization technique which has found applications in many research areas. However, the SA algorithm is limited in its application due to the high computational cost and the difficulties in determining the annealing schedule. This paper demonstrates that the temperature parallel simulated annealing (TPSA) algorithm, a parallel implementation of the SA algorithm, shows great promise to overcome these limitations when applied to continuous functions. The TPSA algorithm greatly reduces the computational time due to its parallel nature, and avoids the determination of the annealing schedule by fixing the temperatures during the annealing process. The main contributions of this paper are threefold. First, this paper explains a simple and effective way to determine the temperatures by applying the concept of critical temperature (T C). Second, this paper presents systematic tests of the TPSA algorithm on various continuous functions, demonstrating comparable performance as well-established sequential SA algorithms. Third, this paper presents results on the study of other parameters important to the TPSA algorithm, including the speedup and the exchange frequency. The results and conclusions presented in this work provide new insights into the further development and expanded applications of the TPSA algorithm.

A parallel simulated annealing algorithm based on functional feature tree modeling for 3D engineering layout design

Computation complexity and modeling complexity are primary clog in 3D engineering layout design. In this paper, a parallel simulated annealing based on multiple Markov chains is employed to deal with the computation complexity. The algorithm increases computing efficiency by allocating computation burden to networked computers. To cope with the modeling complexity, machine components and space limitation are modeled by a hierarchical, feature-based modeling method—the functional feature tree modeling method. And multiple complex design constraints can be described by the representation. In order to support whole layout design solving process, a Computer Aided Layout Design System is introduced briefly. This system not only implements the algorithm and the components representation, but also integrates some other subsystems as evaluation system, constrains editor etc. The methods are demonstrated through a real engineering application, a vehicle engine compartment layout design.

Applications of critical temperature in minimizing functions of continuous variables with simulated annealing algorithm

The simulated annealing (SA) algorithm has been recognized as a powerful technique for minimizing complicated functions. However, a critical disadvantage of the SA algorithm is its high computational cost. Therefore, it is the goal of this paper to investigate the use of the critical temperature in SA to reduce its computational cost. This paper presents a systematic study of the critical temperature and its applications in the minimization of functions of continuous variables with the SA algorithm. Based on this study, a new algorithm was developed to exploit the unique feature of the critical temperature in SA. The new algorithm combines SA and local search to determine global minimum effectively. Extensive tests on a variety of functions demonstrated that the new algorithm provides comparable performance to well-established SA techniques. Furthermore, the new algorithm also improves the determination of the starting temperature for the SA algorithm. The results obtained in this study are expected to be useful for improving the efficiency of SA algorithms, and for facilitating the development of temperature parallel SA algorithms.

Process mining based on parallel recombination simulated annealing algorithm

First,the author described formal description of process mining problem,then proposed a new approach of process mining based on parallel recombination simulated annealing algorithm.This approach adopted causal matrix as the code of process model.In comparison with other similar approaches,it improved on fitness measure function,crossover and mutation genetic operators.Experimental results show that this approach can effectively deal with noise and shorten the mining time.

A multipopulation parallel genetic simulated annealing-based QoS routing and wavelength assignment integration algorithm for multicast in optical networks

In this paper, we propose an integrated Quality of Service (QoS) routing algorithm for optical networks. Given a QoS multicast request and the delay interval specified by users, the proposed algorithm can find a flexible-QoS-based cost suboptimal routing tree. The algorithm first constructs the multicast tree based on the multipopulation parallel genetic simulated annealing algorithm, and then assigns wavelengths to the tree based on the wavelength graph. In the algorithm, routing and wavelength assignment are integrated into a single process. For routing, the objective is to find a cost suboptimal multicast tree. For wavelength assignment, the objective is to minimize the delay of the multicast tree, which is achieved by minimizing the number of wavelength conversion. Thus both the cost of multicast tree and the user QoS satisfaction degree can approach the optimal. Our algorithm also considers load balance. Simulation results show that the proposed algorithm is feasible and effective. We also discuss the practical realization mechanisms of the algorithm.

Reconstruction of periodic unit cells of multimodal random particulate composites using genetic algorithms

We develop a procedure for characterization and reconstruction of periodic unit cells of highly filled, multimodal, particulate composites. Rocpack, a particle packing software, is used to generate the solid propellant microstructures and one- and two-point probability functions are used to describe its statistical morphology. The reconstruction is carried out using a parallel Augmented Simulated Annealing algorithm with a novel mutation operator based on a mass–spring system to eliminate overlap and improve the code performance. Results from the reconstruction procedure, for four-phase random particulate composites of 40–70% packing fraction, are detailed to demonstrate the capabilities of the reconstruction model. The presented results suggest good convergence and repeatability of the optimization scheme, even for high volume fractions, and good scalability of the algorithm.

Optimization of Multiple-Impulse, Multiple-Revolution, Rendezvous- Phasing Maneuvers

A new hybrid optimization approach is proposed for the design of a rendezvous-phasing strategy with combined maneuvers, which is normally considered as a complex multiple-impulse, multiple-revolution, nonlinear rendezvous problem. In this approach, a feasible iteration optimization model is first formulated using a multiple-revolution Lambert algorithm, and a parallel simulated annealing algorithm is employed to locate the unperturbed solution. Subsequently, an infeasible iteration optimization model accounting for trajectory perturbations is formulated, and a sequential quadratic programming algorithm is used to obtain the perturbed solution, with the unperturbed solution as an initial reference solution. The global convergence ability of the proposed approach is verified by solving a classical same-circle rendezvous problem. Two different solutions satisfying Lawden's necessary optimality conditions are located and one solution outperforms an optimal solution previously reported. The proposed approach is further evaluated in a practical two-day rendezvous-phasing mission with different initial conditions. It is shown that this approach is effective and efficient and the combined maneuvers can save propellant at a range of 4-35% when compared with the special-point maneuvers.

Large-Scale 3-D Geometric Reconstruction of the Porcine Coronary Arterial Vasculature Based on Detailed Anatomical Data

The temporal and spatial distribution of coronary blood flow, pressure, and volume are determined by the branching pattern and three-dimensional (3-D) geometry of the coronary vasculature, and by the mechanics of heart wall and vascular tone. Consequently, a realistic simulation of coronary blood flow requires, as a first step, an accurate representation of the coronary vasculature in a 3-D model of the beating heart. In the present study, a large-scale stochastic reconstruction of the asymmetric coronary arterial trees (right coronary artery, RCA; left anterior descending, LAD; and left circumflex, LCx) of the porcine heart has been carried out to set the stage for future hemodynamic analysis. The model spans the entire coronary arterial tree down to the capillary vessels. The 3-D tree structure was reconstructed initially in rectangular slab geometry by means of global geometrical optimization using parallel simulated annealing (SA) algorithm. The SA optimization was subject to constraints prescribed by previously measured morphometric features of the coronary arterial trees. Subsequently, the reconstructed trees were mapped onto a prolate spheroid geometry of the heart. The transformed geometry was determined through least squares minimization of the related changes in both segments lengths and their angular characteristics. Vessel diameters were assigned based on a novel representation of diameter asymmetry along bifurcations. The reconstructed RCA, LAD and LCx arterial trees show qualitative resemblance to native coronary networks, and their morphological statistics are consistent with the measured data. The present model constitutes the first most extensive reconstruction of the entire coronary arterial system which will serve as a geometric foundation for future studies of flow in an anatomically accurate 3-D coronary vascular model.

Development of a parallel optimization method based on genetic simulated annealing algorithm

This paper presents a parallel genetic simulated annealing (PGSA) algorithm that has been developed and applied to optimize continuous problems. In PGSA, the entire population is divided into sub-populations, and in each sub-population the algorithm uses the local search ability of simulated annealing after crossover and mutation. The best individuals of each sub-population are migrated to neighboring ones after a certain number of epochs. An implementation of the algorithm is discussed and the performance is evaluated against a standard set of test functions. PGSA shows some remarkable improvement in comparison with the conventional parallel genetic algorithm and the breeder genetic algorithm (BGA).

A parallel optimization approach for controlling allele diversity in conservation schemes

We propose a novel method to control allelic diversity in conservation schemes based on an optimization problem, characterized by a convex program subject to integer linear constraints. Departing from previous studies considering similar problems, we implement a parallel simulated annealing algorithm to minimize the number of alleles lost across generations. The proposed algorithm shows excellent timing and minimization performances. Execution time decreases linearly with the number of processors used, providing similar results in all cases.

Aerodynamic Shape Optimization Using Computational Fluid Dynamics and Parallel Simulated Annealing Algorithms

Aerodynamic shape design using stochastic optimization methods, such as simulated annealing method to optimize objective functions evaluated by modern state-of-the-art computational fluid dynamics solvers, normally requires enormous computation time to search for the global optimal design. Aerodynamic shape optimization of internal flow systems is studied using Euler/Navier-Stokes solvers and parallel simulated annealing algorithm, which is implemented on parallel computing platforms. A variety of inverse and direct design of internal flow systems are carried out to examine the efficiency and speedup of the parallel simulated annealing algorithms

Parallel Simulated Annealing Algorithms in Global Optimization

Global optimization involves the difficult task of the identification of global extremities of mathematical functions. Such problems are often encountered in practice in various fields, e.g., molecular biology, physics, industrial chemistry. In this work, we develop five different parallel Simulated Annealing (SA) algorithms and compare them on an extensive test bed used previously for the assessment of various solution approaches in global optimization. The parallel SA algorithms consist of various categories: the asynchronous approach where no information is exchanged among parallel runs and the synchronous approaches where solutions are exchanged using genetic operators, or where solutions are transmitted only occasionally, or where highly coupled synchronization is achieved at every iteration. One of these approaches, which occasionally applies partial information exchanges (controlled in terms of solution quality), provides particularly notable results for functions with vast search spaces of up to 400 dimensions. Previous attempts with other approaches, such as sequential SA, adaptive partitioning algorithms and clustering algorithms, to identify the global optima of these functions have failed without exception.

Aerodynamic shape optimization using computational fluid dynamics and parallel simulated annealing algorithms

Aerodynamic shape optimization using computational fluid dynamics and parallel simulated annealing algorithms

PCクラスタ環境における並列シミュレーテッド・アニーリング計算法

PCクラスタ環境における並列シミュレーテッド・アニーリング計算法

Power system network partitioning using tabu search

This paper presents a new algorithm to power system network partitioning based on tabu search (TS). TS is a simple heuristic optimization strategy that can achieve the optimal or near optimal solution within a reasonably short time. A 2-step TS-based algorithm is proposed to divide the system network into several subsystems to optimize the use of parallel computer systems for power system analysis. Test results on two IEEE standard networks are presented and compared with those obtained by other combinatorial optimization approaches, namely: simulated annealing parallel simulated annealing and genetic algorithm. It is shown that the proposed TS-based algorithm has higher efficiency and better convergence than the previously proposed methods.