项目来源
湖北省自然科学基金
项目主持人
贺娟娟
项目受资助机构
武汉科技大学
项目编号
2015CFB335
立项年度
2015
立项时间
未公开
研究期限
未知 / 未知
项目级别
省级
受资助金额
10.00万元
学科
光电子信息
学科代码
未公开
基金类别
未公开
关键词
膜计算 ; 脉冲神经膜系统 ; 膜算法 ; Membrane computing ; Spiking neural P system ; Membrane algorithm
参与者
未公开
参与机构
未公开
项目标书摘要:膜计算模型和由膜计算模型启发得到的膜算法,是目前自然计算领域中最受关注的研究热点之一,也被视作在计算机科学与细胞学的基础上发展起来的一个新颖而极具发展前景的交叉学科。虽然,理论上,膜计算模型具有计算能力强、并行性高、规则简单等优点,但充分利用膜计算模型的理论成果、离散的网状结构信息处理机制、规则的实现技术等仍然有待发展和解决。本研究解决:如何简化膜计算模型使之更易仿真与实现、如何利用膜计算模型中的有向连接与无向连接方式、编码方式、信息传递机制提高算法性能,以及如何把膜算法作为工具在工业生产领域中应用。在计算理论方面,建立了一个仅需6个神经元的通用脉冲神经膜计算模型,改进了目前至少需要10个神经元的结果,减少了实现膜计算模型所需的计算资源,回答了一个公开的重要科学问题:增加组件的数目是否可以减少神经元的数目从而达到图灵通用;发现了在无延迟规则的情况下,同质脉冲神经膜计算模型依然可以保证计算通用性,即具有和图灵机等价的计算能力,简化了实现膜计算模型所需的规则集;在算法设计与应用方面,针对光伏能源利用率有待提高的问题,设计了实时性和全局性兼备的最大功率点跟踪膜算法,在局部遮蔽多干扰情况下,仍能快速跟踪到最大功率点,提高了绿色能源利用率;设计了应用于实际工业生产的改进膜算法,优化了凹印套色系统中的高维参数,将工业制造误差从0.15毫米降低到0.10毫米。本研究提供了需更少资源的通用新型膜计算系统,并为膜计算理论成果推广至实际应用提供方法和理论支持。
Application Abstract: As one of the most popular computing research areas,membrane computing models and algorithms achieved fruitful results,which are developed based on computer science and cytology.The computing devices in membrane computing are called P systems.Membrane-inspired algorithms are a class of parallel intelligent algorithms,which are inspired by the features of P systems.To enrich this research field,several membrane-inspired algorithms are proposed and applied on solving practical problem.In the research,we focus on simplifying P systems and developing potential applications of the membrane algorithm.Theoretically,we study of small spiking neural P systems with cooperating rules and decrease in the number of neurons 10 to 6.Specifically,we construct a Turing universal spiking neural P systems having 6 neurons,which can generate any set of Turing computable natural numbers.We also proved that spiking neural P systems with homogenous neurons and synapses without the feature of delay are also Turing universal.Inspired by memory mechanism in neurons,a ring structure membrane-inspired algorithm with memory mechanism is proposed for tracking the maximum power point of photovoltaic system.The simulation results show that the algorithm performance well that can real-time track the maximum power point even under partial shading condition.Moreover,we proposed a novel membrane algorithm to optimize a set of parameters of industrial roll-to-roll web system.The practical results show that the impact from controllers is reduced under our parameters.This study provides two P systems with less computing resource,and brings important theoretical and practical significance,in terms of promoting the efficiency of membrane systems and algorithms,and its application in industrial field.
项目受资助省
湖北省