A unified programming environment is implemented on our DSS generator, called actDSS, which is a general platform for constructing intelligent specific systems. The architecture of the implemented unified programming ...
详细信息
A unified programming environment is implemented on our DSS generator, called actDSS, which is a general platform for constructing intelligent specific systems. The architecture of the implemented unified programming is a hierarchical connection of a numerical solver and a non-numerical evaluator. The former produces an optimum solution of a problem and the latter evaluates the solution heuristically, controlling the total model that encompasses the former by modifying the constraint and the parameter value. A job assignment problem is used as a demonstration of the architecture, which is formulated as a mixed integer programming problem, In this paper, a general scheme of the unified programming is introduced, Its feasibility on actDSS is demonstrated and finally we discuss how the implementation was possible in the system.
From the formulation point of view, the Integer programming (IF) formulation is no more than indicating some decision variables in a Linear programming model to integer variables. Nevertheless, the interpretation as w...
详细信息
From the formulation point of view, the Integer programming (IF) formulation is no more than indicating some decision variables in a Linear programming model to integer variables. Nevertheless, the interpretation as well as formulation of the IP model is not an easy task at all, because the model has implicit logical meanings in some of the variables and constraints, To explicitly represent the formulation of an integer programming model, eleven logical operators are identified. So the model formulated at this level is called the Logical Integer programming model. To formalize the transformation process of the model to the solvable conventional integer programming model (called Base Level Integer programming model), a series of theorems are derived, This approach is implemented on the system UNIK-IP. UNIK-IP opens a single threaded interface of optimization models with the rule based systems which imply the AND/OR relationships.
This research investigates the automatic identification of typical embedded structures in the Integer programming (IF) models and automatic transformation of the problem to an adequate Lagrangian problem which can pro...
详细信息
This research investigates the automatic identification of typical embedded structures in the Integer programming (IF) models and automatic transformation of the problem to an adequate Lagrangian problem which can provide tight bounds within the acceptable run time, For this purpose, the structural distinctivenesses of variables, constants, blocks of terms, and constraint chunks are identified to specify the structure of the IP model, To assist the identification of the structural distinctiveness, the representation by the knowledge based TP model formulator, UNIK-IP, is adopted, To reason for the structural identification, the hybrid of bottom-up, top-down, and case-based approaches are proposed, A system UNIK-RELAX is developed to implement the approaches proposed in this research.
The paper presents an innovative approach that integrates the arbitrage pricing theory (APT) and artificial neural networks (ANN) to support portfolio management. The integrated approach takes advantage of the synergy...
详细信息
The paper presents an innovative approach that integrates the arbitrage pricing theory (APT) and artificial neural networks (ANN) to support portfolio management. The integrated approach takes advantage of the synergy between APT and ANN in extracting risk factors, predicting the trend of individual risk factor, generating candidate portfolios, and choosing the optimal portfolio. It uses quadratic programming for identifying surrogate portfolios in APT and ANN to predict factor returns. Empirical results indicate that the integrated method beats the benchmark and outperforms the traditional method that uses the ARIMA model.
Modern unified programming models (such as CUDA and SYCL) that combine host (CPU) code and GPU code into the same programming language, same file, and same lexical scope lack adequate support for GPU code specializati...
详细信息
Modern unified programming models (such as CUDA and SYCL) that combine host (CPU) code and GPU code into the same programming language, same file, and same lexical scope lack adequate support for GPU code specialization, which is a key optimization in real-time graphics. Furthermore, current methods used to implement specialization do not translate to a unified environment. In this paper, we create a unified shader programming environment in C++ that provides first-class support for specialization by co-opting C++'s attribute and virtual function features and reimplementing them with alternate semantics to express the services required. By co-opting existing features, we enable programmers to use familiar C++ programming techniques to write host and GPU code together, while still achieving efficient generated C++ and HLSL code via our source-to-source translator.
RNA folding algorithms are challenging dynamic programming tasks to optimize because they are resource-intensive and have a large number of non-uniform dependences. Fortunately, these bioinformatics problems can be re...
详细信息
ISBN:
(纸本)9783030878979;9783030878962
RNA folding algorithms are challenging dynamic programming tasks to optimize because they are resource-intensive and have a large number of non-uniform dependences. Fortunately, these bioinformatics problems can be represented within the polyhedral model. It allows us to use well-known source-to-source optimizing compilers. In this paper, we applied cache efficient strategies developed within the TRACO and Pluto compilers as well as manual codes of transpose and classical loop skewing, and implemented these codes using Intel SDK for OpenCL. This OpenCL platform allows running the unified code on various computing units like multicore CPUs and GPUs including the Intel graphic accelerators. We proposed the Intel OpenCL code generator using parallel implementations. We compared performance on chosen HPC platforms using the basic Nussinov's RNA folding algorithm as an example of non-serial uniform dynamic programming. Experiments were carried out to achieve significant locality improvement and speed-up for cross-platform unified code. We outlined related and future work.
暂无评论