Bi-clustering of gene expression micro array data deals with creating a sub-matrix that shows a high similarity across both genes and conditions. Hi-clustering aims at identifying several bi-clusters that reveal poten...
详细信息
ISBN:
(纸本)9781467365406
Bi-clustering of gene expression micro array data deals with creating a sub-matrix that shows a high similarity across both genes and conditions. Hi-clustering aims at identifying several bi-clusters that reveal potential local patterns from a microarray matrix. In this paper, evolutionary algorithm is used to find bi-clusters of large size which have mean squared residue less than a given threshold, delta. Attention is also given to find bi-clusters with minimum overlapping among themselves by assigning weights to the elements of microarray matrix. Initially, geneticalgorithm (GA) is implemented to derive bi-clusters from microarray matrix. From numerical simulations, it is observed that GA took too much time to converge so as to meet the stopping criteria. To further improve the performance of GA, parallel GA (PGA) is implemented with an objective, so as to efficiently handle the problem of slow convergence encountered in traditional GA. A framework of Coarse grained parallel genetic algorithm (CgPGA) for bi-clustering is implemented in this paper. The results obtained from CgPGA are quite encouraging as CgPGA took very less time to meet the stopping criteria. The bi-clusters derived by CgPGA are larger in size, which is one of the primary objective of bi-clustering problem. The experiment was performed on micro array dataset i.e. yeast Saccharomyces cerevisiae cell cycle.
作者:
Potuzak, TomasUniv W Bohemia
Fac Sci Appl Dept Comp Sci & Engn NTIS European Ctr Excellence Plzen 30614 Czech Republic
In this paper, we explore the features of the sparsely synchronized parallel genetic algorithm for the road traffic network division. The algorithm is an alternative to a commonly used island model for the paralleliza...
详细信息
ISBN:
(纸本)9781467369367
In this paper, we explore the features of the sparsely synchronized parallel genetic algorithm for the road traffic network division. The algorithm is an alternative to a commonly used island model for the parallelization of the geneticalgorithms. The algorithm employs the parallelization of particular phases of the geneticalgorithm (fitness values calculation, crossover, etc.). However, the threads of the geneticalgorithm are not synchronized in every generation, but rather only once per several generations or even not at all. The lack of the synchronization leads to the inconsistencies in the shared memory, which does not have to be a problem considering the stochastic nature of the geneticalgorithms. The investigation of the features and usability of the sparse synchronization of the parallel genetic algorithm (with application for the road traffic network division) is the main theme of this paper.
To make deep neural networks automatically achieve the same or better performance compared with those in hand-optimized libraries, Tensor Virtual Machine (TVM) has combined a geneticalgorithm (GA) with its AutoTVM au...
详细信息
To make deep neural networks automatically achieve the same or better performance compared with those in hand-optimized libraries, Tensor Virtual Machine (TVM) has combined a geneticalgorithm (GA) with its AutoTVM auto-tuning process. The geneticalgorithm of TVM has a primary and classic design, with restrictions in terms of searching scope, ability, and efficiency. Meanwhile, the current AutoTVM process is time-consuming. The whole process may last hours on GPUs. As such, we propose a new auto-tuning method that is based on a parallel GA and takes advantage of the strengths of the Roofline model-based cost models and machine learning classification models to widen the search scope and improve search efficiency. The new auto-tuning method achieves double optimization on both tuning results and tuning time. A series of experiments show that the new way improves the inference time of typical deep networks by about 8-14% and speeds up the time consumption of the auto-tuning process up to 1.2-1.52x on GPUs compared with the original GA process of AutoTVM.
作者:
Akopov, Andranik S.Natl Res Univ
Fac Business Informat Higher Sch Econ Business Analyt Dept Kirpichnaya Str 33 Moscow 105187 Russia
This work presents a novel approach to designing the parallel genetic algorithm (GA) with fading selection for the solving of the problem of the shareholder value maximisation of an oil company. The algorithm based on...
详细信息
This work presents a novel approach to designing the parallel genetic algorithm (GA) with fading selection for the solving of the problem of the shareholder value maximisation of an oil company. The algorithm based on the dynamical interaction of synchronised processes, which are interdependent GAs having own separate evolutions of their populations. The developed system allows users to find a set of non-dominated investment projects (Pareto-efficient solutions), allowing to pick a particular solution in accordance with their utility function.
parallel versions of a geneticalgorithm based on the hybrid MPI-OpenMP model are implemented to optimize circulant networks, which are of practical interest in the design of supercomputer systems and systems on a chi...
详细信息
parallel versions of a geneticalgorithm based on the hybrid MPI-OpenMP model are implemented to optimize circulant networks, which are of practical interest in the design of supercomputer systems and systems on a chip. An analysis of the efficiency of parallel programs with different numbers of MPI processes and OpenMP threads on a cluster of Kunpeng processors has been carried out. The speed-up of several hybrid parallel computing schemes was experimentally evaluated and analyzed. Two bottlenecks in terms of efficiency in parallel execution of the algorithm are identified and methods for their solution are proposed. By means of the parallel genetic algorithm the descriptions of circulant networks with better average distance and bisection width for the known large circulant networks were obtained.
The test task scheduling problem (TTSP) has attracted increasing attention due to the wide range of automatic test systems applications, despite the fact that it is an NP-complete problem. The main feature of TTSP is ...
详细信息
The test task scheduling problem (TTSP) has attracted increasing attention due to the wide range of automatic test systems applications, despite the fact that it is an NP-complete problem. The main feature of TTSP is the close interactions between task sequence and the scheme choice. Based on this point, the parallel implantation of geneticalgorithm, called parallel genetic algorithm (PGA), is proposed to determine the optimal solutions. Two branches-the tasks sequence and scheme choice run the classic geneticalgorithm independently and they balance each other due to their interaction in the given problem. To match the frame of the PGA, a vector group encoding method is provided. In addition, the fitness distance coefficient (FDC) is first applied as the measurable step of landscape to analyze TTSP and guide the design of PGA when solving the TTSP. The FDC is the director of the search space of the TTSP, and the search space determinates the performance of PGA. The FDC analysis shows that the TTSP owes a large number of local optima. Strong space search ability is needed to solve TTSP better. To make PGA more suitable to solve TTSP, three crossover and four selection operations are adopted to find the best combination. The experiments show that due to the characteristic of TTSP and the randomness of the algorithm, the PGA has a low probability for optimizing the TTSP, but PGA with Nabel crossover and stochastic tournament selection performs best. The assumptions of FDC are consistent with the success rate of PGA when solving the TTSP.
We introduce a method for imaging the earthquake source dynamics from the inversion of ground motion records based on a parallel genetic algorithm. The source model follows an elliptical patch approach and uses the st...
详细信息
We introduce a method for imaging the earthquake source dynamics from the inversion of ground motion records based on a parallel genetic algorithm. The source model follows an elliptical patch approach and uses the staggered-grid split-node method to simulate the earthquake dynamics. A statistical analysis is used to estimate errors in both inverted and derived source parameters. Synthetic inversion tests reveal that the average rupture speed (V-r), the rupture area, and the stress drop () may be determined with formal errors of similar to 30%, similar to 12%, and similar to 10%, respectively. In contrast, derived parameters such as the radiated energy (E-r), the radiation efficiency ((r)), and the fracture energy (G) have larger errors, around similar to 70%, similar to 40%, and similar to 25%, respectively. We applied the method to the M-w 6.5 intermediate-depth (62km) normal-faulting earthquake of 11 December 2011 in Guerrero, Mexico. Inferred values of =29.26.2MPa and (r)=0.260.1 are significantly higher and lower, respectively, than those of typical subduction thrust events. Fracture energy is large so that more than 73% of the available potential energy for the dynamic process of faulting was deposited in the focal region (i.e., G=(14.43.5)x10(14)J), producing a slow rupture process (V-r/V-S=0.470.09) despite the relatively high energy radiation (E-r=(0.54 +/- 0.31)x10(15)J) and energy-moment ratio (E-r/M-0=5.7x10(-5)). It is interesting to point out that such a slow and inefficient rupture along with the large stress drop in a small focal region are features also observed in both the 1994 deep Bolivian earthquake and the seismicity of the intermediate-depth Bucaramanga nest.
parallel genetic algorithms have proved to be a successful method for solving the protein folding problem. In this paper we propose a simple geneticalgorithm with optimum population size, mutation rate and selection ...
详细信息
ISBN:
(纸本)9783319031071
parallel genetic algorithms have proved to be a successful method for solving the protein folding problem. In this paper we propose a simple geneticalgorithm with optimum population size, mutation rate and selection strategy which is parallelized with MapReduce architecture for finding the optimal conformation of a protein using the two dimensional square HP model. We have used an enhanced framework for map Reduce which increased the performance of the private clouds in distributed environment. The proposed geneticalgorithm was tested several bench mark of synthetic sequences. The result shows that GA converges to the optimum state faster than the traditional
geneticalgorithms (GA) are widely used in the literature to extract interesting association rules. However, they are time consuming mainly due to the growing size of databases. To speed up this process, we propose tw...
详细信息
geneticalgorithms (GA) are widely used in the literature to extract interesting association rules. However, they are time consuming mainly due to the growing size of databases. To speed up this process, we propose two parallel GAs (ARMGPU and ARM-CPU/GPU). In ARM-GPU, parallelism is used to compute the fitness which is the most time consuming task;while, ARM-CPU/GPU proposes a two-level-based parallel GA. In the first level, the different cores of the CPU execute a GAARM on a sub-population. The second level of parallelism is used to compute the fitness, in parallel, on GPU. To validate the proposed two parallel GAs, several tests were conducted to solve well-known large ARM instances. Obtained results show that our parallelalgorithms outperform state-of-the-art exact algorithms (APRIORI and FP-GROWTH) and approximate algorithms (SEGPU and ME-GPU) in terms of execution time.
Recently, heat-integrated pressure-swing distillation (HIPSD) has been explored for recovering isopropanol and benzene from wastewater, but the process remains highly energy-intensive. Given the dilute nature of the f...
详细信息
Recently, heat-integrated pressure-swing distillation (HIPSD) has been explored for recovering isopropanol and benzene from wastewater, but the process remains highly energy-intensive. Given the dilute nature of the feed, intensified extractive distillation with an integrated preconcentration column (IED) is more suitable. However, previous researches have often overlooked the critical role of pressure and lacked rigorous optimization of operating pressure in such systems. In this article, we developed novel extractive pressure-swing distillation with integrated feed preconcentration/solvent recovery column (IEPSD) by introducing a pressure-swing configuration and incorporate energy-saving technologies to achieve more sustainable and cost-effective separation processes. First, thermodynamic analysis is conducted to explore the effect of pressure on extractive pressure-swing distillation. Then, a parallel genetic algorithm is applied for rigorous optimization, followed by the application of heat integration and heat pump. The IEPSD is superior to IED in terms of total annual cost (TAC) and CO2 emission. With the inclusion of energy-saving technologies, IEDHI and IEPSDHI further reduced TAC by 5.29% and 7.40%, and cut CO2 emissions by 25.18% and 24.03%, respectively, compared to their base processes. The use of a heat pump is particularly advantageous for IEPSD due to its lower boiling temperatures under vacuum pressure, leading to the development of IEPSDHIHP, which offered an additional 22.36% CO2 reduction and marginally lower TAC than IEPSDHI. Ultimately, IEPSDHIHP proves to be the best option, offering a 51.15% TAC reduction and 68.42% CO2 emissions reduction compared to HIPSD. In summary, the developed IED and IEPSD processes, especially with heat integration and heat pump technologies, offer significant economic and environmental advantages over conventional HIPSD.
暂无评论