算法隐含并行性的物理模型

基金项目:

国家自然科学基金(60702075);中国博士后科学基金(2007410385);四川省教育厅自然科学重点项目(07ZA014)

详细信息

Physical Model of Implicit Parallelism in Algorithms

1.
School of Computer Science and Engineering,University of Electronic Science and Technology of China Chengdu 610054;
2.
Parallel Computing Laboratory,Chengdu University of Information Technology Chengdu 610225

摘要: 利用物理学原理对算法的隐含并行性进行了分析，提出算法的不确定性和高熵态是隐含并行性出现的根源，但算法的隐含并行性会导致算法结果的不确定性。智能算法中先验知识确定程序的搜索方向，隐含并行性提供了对解空间的高速并行搜索，为研究和设计智能算法提供了理论基础。提高算法的先验信息量和隐含并行能力可以得到具备较高智能水平的算法。
- 先验知识 /
- 熵 /
- 高熵态 /
- 隐含并行性 /
- 不确定性
Abstract: According to the analysis of the implicit parallelism in algorithms, a physical model of implicit parallelism in algorithm is proposed. The uncertainty and high entropy state of algorithm bring forward implicit parallelism, but the implicit parallelism of algorithms may cause the uncertainty of algorithm results. In the proposed model, the search direction in result space is decided by the apriori knowledge and the high speed parallel search ability is decided by the implicit parallelism. The implicit parallelism of genetic algorithm and simulated annealing algorithm are analyzed by this model.
- apriori knowledge /
- entropy /
- high entropy state /
- implicit parallelism /
- uncertainty