This paper investigates a hybrid algorithm which utilizes exact and heuristic methods to optimise asset selection and capital allocation in portfolio optimisation. The proposed method is composed of a customised population based incremental learning procedure and a mathematical programming application. It is based on the standard Markowitz model with additional practical constraints such as cardinality on the number of assets and quantity of the allocated capital. Computational experiments have been conducted and analysis has demonstrated the performance and effectiveness of the proposed approach.
{"title":"A Population-Based Incremental Learning Method for Constrained Portfolio Optimisation","authors":"Yan Jin, R. Qu, J. Atkin","doi":"10.1109/SYNASC.2014.36","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.36","url":null,"abstract":"This paper investigates a hybrid algorithm which utilizes exact and heuristic methods to optimise asset selection and capital allocation in portfolio optimisation. The proposed method is composed of a customised population based incremental learning procedure and a mathematical programming application. It is based on the standard Markowitz model with additional practical constraints such as cardinality on the number of assets and quantity of the allocated capital. Computational experiments have been conducted and analysis has demonstrated the performance and effectiveness of the proposed approach.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129423981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Modern software and hardware designs are mostly hierarchical. Moreover, while the design specification is defined up-down, the design implementation and verification are done down-up. In such a case, as a rule, coverage properties for simulation-based verification are defined inconsistently for different stages of the design flow. The fact leads to the well known explosion of bug rate, when we pass from the lower design stage to the upper one. In this paper, we propose a new approach that allows propagation of quantitative properties from the upper stage of the design flow to the lower ones and their incremental computation on the components. The approach may be applied to any design, modeled as a Finite State Machine (FSM) or other formalisms, which eventually lead to weighted labeled trees. We use Weighted Monadic-Second Order Logic (WMSOL) to describe the desired families of quantitative properties and sum-like weighted labeled trees to describe the decomposition of the behaviour of the FSM. The last notion is based on a generalization of disjoint unions of structures with additional links between the components. Our approach shows how computation of a quantitative property, definable as a WMSOL formula on the upper stage of the design may be reduced for certain cost to incremental computations of effectively derivable WMSOL-definable properties on the components. We provide several examples of families of such properties and discuss different aspects, related to the applicability of our approach. The approach is new and provides a uniform theoretical basis for analyzing WMSOL-definable properties on hierarchical structures.
{"title":"Analyzing WMSOL Definable Properties on Sum-Like Weighted Labeled Trees","authors":"E. Ravve","doi":"10.1109/SYNASC.2014.57","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.57","url":null,"abstract":"Modern software and hardware designs are mostly hierarchical. Moreover, while the design specification is defined up-down, the design implementation and verification are done down-up. In such a case, as a rule, coverage properties for simulation-based verification are defined inconsistently for different stages of the design flow. The fact leads to the well known explosion of bug rate, when we pass from the lower design stage to the upper one. In this paper, we propose a new approach that allows propagation of quantitative properties from the upper stage of the design flow to the lower ones and their incremental computation on the components. The approach may be applied to any design, modeled as a Finite State Machine (FSM) or other formalisms, which eventually lead to weighted labeled trees. We use Weighted Monadic-Second Order Logic (WMSOL) to describe the desired families of quantitative properties and sum-like weighted labeled trees to describe the decomposition of the behaviour of the FSM. The last notion is based on a generalization of disjoint unions of structures with additional links between the components. Our approach shows how computation of a quantitative property, definable as a WMSOL formula on the upper stage of the design may be reduced for certain cost to incremental computations of effectively derivable WMSOL-definable properties on the components. We provide several examples of families of such properties and discuss different aspects, related to the applicability of our approach. The approach is new and provides a uniform theoretical basis for analyzing WMSOL-definable properties on hierarchical structures.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130680470","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Huge data advent in high-performance computing (HPC) applications such as fluid flow simulations usually hinders the interactive processing and exploration of simulation results. Such an interactive data exploration not only allows scientiest to 'play' with their data but also to visualise huge (distributed) data sets in both an efficient and easy way. Therefore, we propose an HPC data exploration service based on a sliding window concept, that enables researches to access remote data (available on a supercomputer or cluster) during simulation runtime without exceeding any bandwidth limitations between the HPC back-end and the user front-end.
{"title":"Interactive Data Exploration for High-Performance Fluid Flow Computations through Porous Media","authors":"N. Perovic, J. Frisch, R. Mundani, E. Rank","doi":"10.1109/SYNASC.2014.68","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.68","url":null,"abstract":"Huge data advent in high-performance computing (HPC) applications such as fluid flow simulations usually hinders the interactive processing and exploration of simulation results. Such an interactive data exploration not only allows scientiest to 'play' with their data but also to visualise huge (distributed) data sets in both an efficient and easy way. Therefore, we propose an HPC data exploration service based on a sliding window concept, that enables researches to access remote data (available on a supercomputer or cluster) during simulation runtime without exceeding any bandwidth limitations between the HPC back-end and the user front-end.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127375489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper is concerned on analytical treatment of non-linear differential equation of a viscous boundary layer flow due to a moving sheet. An analytic approximate technique, namely Optimal Homotopy Asymptotic Method (OHAM) is employed into a new version for this purpose. It is proved that OHAM provide accurate solution for the nonlinear differential equation of the third-order with initial and boundary conditions. Our procedure provides us with a convenient way to optimally control the convergence of the solution, such that the accuracy is always guaranteed. An excellent agreement of the approximate solution with the numerical results has been demonstrated. This work shows the general validity and the great potential of the OHAM for solving strongly nonlinear differential equation.
{"title":"Optimal Homotopy Asymptotic Method for Viscous Boundary Layer Flow in Unbounded Domain","authors":"R. Ene, V. Marinca, R. Negrea","doi":"10.1109/SYNASC.2014.22","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.22","url":null,"abstract":"This paper is concerned on analytical treatment of non-linear differential equation of a viscous boundary layer flow due to a moving sheet. An analytic approximate technique, namely Optimal Homotopy Asymptotic Method (OHAM) is employed into a new version for this purpose. It is proved that OHAM provide accurate solution for the nonlinear differential equation of the third-order with initial and boundary conditions. Our procedure provides us with a convenient way to optimally control the convergence of the solution, such that the accuracy is always guaranteed. An excellent agreement of the approximate solution with the numerical results has been demonstrated. This work shows the general validity and the great potential of the OHAM for solving strongly nonlinear differential equation.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128004973","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
For most applications, taking full advantage of the memory system is key to achieving good performance on GPUs. In this paper, we introduce register caching, a novel idea where registers of multiple threads are combined and used as a shared, last level, manually managed cache for the contributing threads. This method is enabled by the shuffle instruction recently introduced in Nvidia's Kepler GPU architecture, which allows threads in the same warp to exchange data directly, previously only possible by going through shared memory. We evaluate our proposal with a stencil computation benchmark, achieving speedups of up to 2.04, compared to using shared memory on a GTX680 GPU. Stencil computations form the core of many scientific applications, which can therefore benefit from our proposal. Furthermore, our method is not limited to stencil computations, but is applicable to any application with a predictable memory access pattern suitable for manual caching.
{"title":"Register Caching for Stencil Computations on GPUs","authors":"Thomas L. Falch, A. Elster","doi":"10.1109/SYNASC.2014.70","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.70","url":null,"abstract":"For most applications, taking full advantage of the memory system is key to achieving good performance on GPUs. In this paper, we introduce register caching, a novel idea where registers of multiple threads are combined and used as a shared, last level, manually managed cache for the contributing threads. This method is enabled by the shuffle instruction recently introduced in Nvidia's Kepler GPU architecture, which allows threads in the same warp to exchange data directly, previously only possible by going through shared memory. We evaluate our proposal with a stencil computation benchmark, achieving speedups of up to 2.04, compared to using shared memory on a GTX680 GPU. Stencil computations form the core of many scientific applications, which can therefore benefit from our proposal. Furthermore, our method is not limited to stencil computations, but is applicable to any application with a predictable memory access pattern suitable for manual caching.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128515461","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
One of the main problems practitioners have to deal with is the identification of change and defect proneness of source code entities (e.g., Classes). During the last years a lot of techniques have been employed for predicting change and defect proneness of classes. In this paper we study the capabilities of Genetic Programming for performing the addressed problem by measuring the precision and recall of the obtained predictions.
{"title":"How Good Is Genetic Programming at Predicting Changes and Defects?","authors":"C. Marinescu","doi":"10.1109/SYNASC.2014.78","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.78","url":null,"abstract":"One of the main problems practitioners have to deal with is the identification of change and defect proneness of source code entities (e.g., Classes). During the last years a lot of techniques have been employed for predicting change and defect proneness of classes. In this paper we study the capabilities of Genetic Programming for performing the addressed problem by measuring the precision and recall of the obtained predictions.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"2 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116812239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Hansen coefficients are one of the most important tools in the analytical or semi-analytical methods of celestial mechanics. The problem of efficient computation was investigated in detail in many papers, but this subject is still open and to date there are not standard algorithms to generate the Hansen coefficients with accuracy and good computation time. Although the analytical expressions are known through their accuracy, the formulation by recurrent relations was preferred due to their stability and applicability in the case when a large number of the Hansen coefficients are needed. In this paper we turn our attention to the polynomial approximations and we present a method to generate the polynomials associated to the Hansen coefficients and their derivatives in the particular case used in expression of disturbing function due to central-body. Using the recurrent formulas corresponding to this particular case, we define difference equations and, by their iteration, we generate the polynomials. By computational tests, we compare our polynomial approximations for the Hansen coefficients and their derivatives with the recurrent solutions in order to prove the non-regression of the accuracy and the gain in computation time.
{"title":"On the Computation of the Hansen Coefficients","authors":"Petru Calin Bazavan, Lucian Barbulescu, P. Cefola","doi":"10.1109/SYNASC.2014.24","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.24","url":null,"abstract":"The Hansen coefficients are one of the most important tools in the analytical or semi-analytical methods of celestial mechanics. The problem of efficient computation was investigated in detail in many papers, but this subject is still open and to date there are not standard algorithms to generate the Hansen coefficients with accuracy and good computation time. Although the analytical expressions are known through their accuracy, the formulation by recurrent relations was preferred due to their stability and applicability in the case when a large number of the Hansen coefficients are needed. In this paper we turn our attention to the polynomial approximations and we present a method to generate the polynomials associated to the Hansen coefficients and their derivatives in the particular case used in expression of disturbing function due to central-body. Using the recurrent formulas corresponding to this particular case, we define difference equations and, by their iteration, we generate the polynomials. By computational tests, we compare our polynomial approximations for the Hansen coefficients and their derivatives with the recurrent solutions in order to prove the non-regression of the accuracy and the gain in computation time.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126842860","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Multi-objective Optimization Evolutionary Algorithms are widely employed for solving different real-world optimization problems. Usually their runs involve a considerable amount of time because of the need to evaluate many functions. This particularity makes them good candidates of parallelization. In this work we investigate the benefits of the GPU implementation of the Non-dominated Sorting Genetic Algorithm II (NSGA-II) versus its CPU implementation in terms of the execution time.
{"title":"NSGA-II: Implementation and Performance Metrics Extraction for CPU and GPU","authors":"Florina Roxana Padurariu, C. Marinescu","doi":"10.1109/SYNASC.2014.72","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.72","url":null,"abstract":"Multi-objective Optimization Evolutionary Algorithms are widely employed for solving different real-world optimization problems. Usually their runs involve a considerable amount of time because of the need to evaluate many functions. This particularity makes them good candidates of parallelization. In this work we investigate the benefits of the GPU implementation of the Non-dominated Sorting Genetic Algorithm II (NSGA-II) versus its CPU implementation in terms of the execution time.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125994966","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Even if clouds are not fit for high-end HPC applications, they could be profitably used to bring the power of economic and scalable parallel computing to the masses. But this requires both simple development environments, able to exploit cloud scalability, and the capability to easily predict the cost of HPC application runs. This paper presents a framework built on the top of a cloud aware programming platform (mOSAIC) for the development of bag-of-tasks scientific applications. The framework integrates a cloud-based simulation environment able to predict the behavior of the developed applications. Simulations enable the developer to predict at an early development stage performance and cloud resource usage, and so the infrastructure lease cost on a public cloud. The paper sketches the framework organization and discusses the approach followed for application development. Moreover, some validation tests of prediction results are presented.
{"title":"Early Prediction of the Cost of HPC Application Execution in the Cloud","authors":"M. Rak, Mauro Turtur, Umberto Villano","doi":"10.1109/SYNASC.2014.61","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.61","url":null,"abstract":"Even if clouds are not fit for high-end HPC applications, they could be profitably used to bring the power of economic and scalable parallel computing to the masses. But this requires both simple development environments, able to exploit cloud scalability, and the capability to easily predict the cost of HPC application runs. This paper presents a framework built on the top of a cloud aware programming platform (mOSAIC) for the development of bag-of-tasks scientific applications. The framework integrates a cloud-based simulation environment able to predict the behavior of the developed applications. Simulations enable the developer to predict at an early development stage performance and cloud resource usage, and so the infrastructure lease cost on a public cloud. The paper sketches the framework organization and discusses the approach followed for application development. Moreover, some validation tests of prediction results are presented.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127348604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We present an algebraic framework to represent indefinite nested sums over hyper geometric expressions in difference rings. In order to accomplish this task, parts of Karr's difference field theory have been extended to a ring theory in which also the alternating sign can be expressed. The underlying machinery relies on algorithms that compute all solutions of a given parameterized telescoping equation. As a consequence, we can solve the telescoping and creative telescoping problem in such difference rings.
{"title":"A Streamlined Difference Ring Theory: Indefinite Nested Sums, the Alternating Sign, and the Parameterized Telescoping Problem","authors":"Carsten Schneider","doi":"10.1109/SYNASC.2014.12","DOIUrl":"https://doi.org/10.1109/SYNASC.2014.12","url":null,"abstract":"We present an algebraic framework to represent indefinite nested sums over hyper geometric expressions in difference rings. In order to accomplish this task, parts of Karr's difference field theory have been extended to a ring theory in which also the alternating sign can be expressed. The underlying machinery relies on algorithms that compute all solutions of a given parameterized telescoping equation. As a consequence, we can solve the telescoping and creative telescoping problem in such difference rings.","PeriodicalId":150575,"journal":{"name":"2014 16th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132460158","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：