A Study on Integrated of Chaotic Dinamical Memory

混沌动态记忆整合研究

• 批准号: 09044196
• 负责人: YAMAKAWA Takeshi
• 金额: $ 5.18万
• 依托单位: Fuzzy Logic Systems Institute
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for international Scientific Research
• 财政年份: 1997
• 资助国家: 日本
• 起止时间: 1997 至 1998
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-09044196/
• 关键词: chaos unit; saturated N-shaped nonlinear function; saturated inverse-N shaped nonlinear function; chaotic itinerancy; non-equilibrium cross-coupled network; chaos memory; global minimum; Traveling Salesman Problem (TSP); パターン想起; 非平衡相互結合型ネットワーク

The possibility of an associative memory, which is constructed by cross-coupling the chaos units, is discussed considering silicon implementation. The following conclusions are obtained from this research.1. Design of a Chaos Unit : The chaos unit is constructed with a nonlinear device and a delay element.(1) It is a saturated inverse-N-shaped nonlinear function that is suitable for realizing effective chaotic itinerancy.(2) It is a saturated N-shaped nonlinear function that is suitable for solving Traveling Salesman Problem. The network employing this nonlinear devices guarantees the global minimum in obtaining the solution by increasing the gains of weighting factors.(3) The internal value of each chaos unit, the weighted sum of inputs, can be the detectable measure of similarity to design the network.2. Suitability for Silicon Implementation(1) The chaos unit of interest exhibits simplicity and thus is suitable for an integrated circuit.(2) The chaos network obtained in this research is insensitive to the variance of device dimensions to be stable in dynamical behavior and thus is suitable for an integrated circuit.

考虑到硅实现，讨论了通过交叉耦合混沌单元构建关联存储器的可能性。本研究得到以下结论： 1．混沌单元的设计：混沌单元由非线性器件和延迟元件构成。(1)它是一个饱和的逆N型非线性函数，适合于实现有效的混沌迭代。(2)它是一个饱和的N型非线性函数，适合于求解旅行商问题。采用这种非线性器件的网络通过增加权重因子的增益来保证获得解的全局最小值。(3)每个混沌单元的内部值，即输入的加权和，可以作为设计网络的相似性的可检测度量。2．硅实现的适用性（1）感兴趣的混沌单元表现出简单性，因此适用于集成电路。（2）本研究中获得的混沌网络对器件尺寸的变化不敏感，动态行为稳定，因此适用于集成电路。

项目成果

Nikola K.Kasabov: "A Fuzzy Neural Network for Knowledge Acquisition in Complex Time Series" Control and Cybernetics. 25巻・4号. 1-19 (1998)

Nikola K. Kasabov：“复杂时间序列中知识获取的模糊神经网络”控制与控制论，第 25 卷，第 4 期。1-19 (1998)。

Tomasz J.Cholewo: "Sequential Network Construction for Time Series Prediction" Proceedings of 1997 International Conference on Neural Networks (ICNN '97). Vol.4. (1997)

Tomasz J.Cholewo：“时间序列预测的序列网络构建”1997 年国际神经网络会议 (ICNN 97) 会议记录。

Harold Szu: "Advanced Neural Nets for Visual Image Communication" Proceedings of the 5th International Conference on Soft Computing and Information/Intelligent Ststems (IIZUKA '99)/Methodologies for the Conception, Design and Application of Soft Computing

Harold Szu：“用于视觉图像通信的高级神经网络”第五届软计算和信息/智能系统国际会议（IIZUKA 99）/软计算概念、设计和应用方法论的论文集

Arno Joentgen: "Dynamic Data Analysis : Similarity Between Trajectories" Proceedings of Fuzzy-Neuro Systems '98. 98-105 (1998)

Arno Joentgen：“动态数据分析：轨迹之间的相似性”模糊神经系统 98 论文集。

Masayoshi Shimono: "Design and Analysis of a Non-equilibrium Cross-Coupled Network with a Detectable Similarity Measure" IEEE Transactions on Neural Networks. (to be published).

Masayoshi Shimono：“具有可检测相似性度量的非平衡交叉耦合网络的设计和分析”IEEE Transactions on Neural Networks。