2，知识作为一个与控制程序分开的实体进行编码和维护。因此，它不与控制程序一起编译，这就允许增量式地增加和修改(改进）知识库，而不需重新编译控制程序。而且， 在某些情况下有可能就同一些控制程序使用不同的知识库来产生不同类型的专家系统, 这样一些系统称为专家系统外壳，因为它们可以装有不同的知识库。 Expert systems are capable of explaining how a particular conclusion was reached,and why requested information is needed during a consultation. This is important as it gives the user a chance to assess and understand the system's reasoning ability, thereby improving the user's confidence in the system.
3，专家系统能够解释一个,特定的结论是怎么得到的，以及在咨询过程中为什么需要所要求的信息。这是重要的，因为它给用户一种机会来评价和理解专家系统的推理能力， 因此提高了用户对系统的信任。 4.Expert systems use symbolic representations for knowledge (rules,networks,or frames) and perform their inference through symbolic computations that closely resemble manipulations of natural language. (An exception to this is the expert system based on neural network architectures. )
4.专家系统对知识使用符号表示法（规则、网络或框架：)，并且通过符号计算进行推理，这非常类似于自然语言处理（例外是基于神经网络结构的专家系统）。
5.Expert systems often reason with metaknowledge; that is, they reason with knowledge about themselves,and their own knowledge limits and capabilities.
5.专家系统经常用元知识进行推理，即它们用有关它们自身的知识以及它们自己的知识范围和能力来推理。
The Reality and Future of Expert Systems
专家系统的现状与未来 Companies will not truly use expert systems (ES) for business operations until ES applieations can be fully integrated with other conventional software and data bases,can be developed and operated cost-effectively,and can be validated to protect against errors that can have a deleterious impact on business operations'-4-1. The AI industry is making significant efforts toward defining software standards and development methodology,including verification and validation techniques—the ES technology issues identified by CI-Os as major roadblocks hindering ES progress.
各公司只有当专家系统应用能完全与其他常规软件和数据库融合为一体、能被成本有效地开发和运作，并证实能防止对业务运作产生有害影响的错误时，他们才将真正地把专家系统应用于业务经营中。人工智能业正在致力于制定软件标准和开发方法论，包括验证与确认技术，专家系统的这些技术问题被信息部门主管视作妨碍专家系统取得进展的主要障碍。
The two problems with the current knowledge encoding process are current ES techniques do not facilitate modeling of commonsense knowledge, and the knowledge base does not learn from its experience. Case-based reasoning (CBR) and the technologies that are emerging from the CYC project at MCC Corp , address both of these problems.
当前知识编码过程的两个问题是：(1)当前的专家系统技术不便于常识性知识的建模；知识库不能从经验中学习。基于实例的推理（CBR）技术和从MCC公司的CYC项目中出现的技术都强调了这两个问题。
CBR is a theory that deals with memory,learning,planning,and problem solving. It assumes that the human expert derives knowledge from experience (called cases) and can better articulate knowledge as experience than as rulesC6]. CBR has been modeled in new AI tools to improve their problem-solving techniques and to enable them to adapt to new situations. A case is described to the computer in terms of its features. Determining the appropriate case features is
the main task of knowledge engineering. CBR does not capture commonsense knowledge that a human expert may use in solving a problem. NCC Corp. seeks to resolve this issue by developing an extremely large knowledge base f commonsense. Its CYC project has developed the following special tools for efficient-y capturing and processing commonsense knowledge:
CBR是一种处理记忆、学习、计划和问题求解的理论。它假设人类专家是从经验（称做实例）获得知识的，而且经验比规则能更好地明确表达知识。CBR已按一些新的人工智能工具建模，以改进它们的问题求解技术和使它们适应各种新的情况。实例是从其特征出发描述给计算机的。确定合适的实例特殊是知识工程的主要任务。CBR不收集人类专家解决问题时可能使用的常识性知识。NCC公司寻求通过开发—个非常大的,常识知识库来解决这个问题。该公司的CYC项目已开发了高效获取和处理常识性知识的下列专门工具：
Representation language, CYCL, which handles prepositional attitudes, such as beliefs,goals,and purpose;
? 表示法语言CYCL，它处理诸如信念、目标和目的等这样的前法。
knowledge base, which exists at two levels: (1)easy user and external system interacting, (2) efficient processings facility for automatically translating sentences from the user level into the most appropriate representation for processing。
? 知识库，它分两个层次：（1）简便的用户和外部系统交互层；（2）高效处理层、一
个把用户层的句子自动翻译成最适于处理的表示的设施。
Dealing with imprecision has been a problem for AI. And inference mechanism that uses fuzzy logic provides a way to process imprecise rules and data easily and efficiently. Fuzzy logic performs an approximation of possible outcomes, clusters these outcomes as overlapping gradations,and processes these as series of overlapping sets.
对人工智能来说，处理不精确性一直是个问题。应用模糊逻辑的推理机制为简便且有效地处理不精确的规则和数据提供了一条途径。模糊逻辑执行可能结果的近似,把这些结果群集成一些重叠的层次，再把它们作为重叠集合序列进行处理。
The architecture of the next generation of expert systems provides for full integration of expert systems with external systems and data bases. Although AI researchers are addressing most of the technological limitations of early expert systems, companies are still waiting for software standards and development methodologies. The success of neural networks,CASE technology,and client/server computing in the last two years indicates that US companies realize they must modernize their information systems. The new generation of information systems will most likely use AI and other state-of-the-art technologies.
下―代专家系统的体系结构为专家系统同外部系统和数据库的完全集成做准备。虽然人工智能研究人员正在致力于克服早期专家系统的大部分技术限制，但各公司仍在等待软件标准和开发方法论。在过去两年中，神经网络、计算机辅助软件工程（CASE）技术和客户机-服务器计算的成功表明，美国一些公司认识到它们必须使其信息系统现代化。新一代的信息处理系统将最可能使用人工智能和其他一些当前最新技术。
Unit 10 Data Structure and Algorithms
数据结构和算法
10.1 Abstract Data Types and Algorithms
抽象数据类型和算法
Data abstraction is a central concept in program design. The abstraction defines the domain and structure of the data.along with a collection of operations that access the data. The abstraction,called an abstract data type (ADT) ,creates a user defined data type whose operations specify how a client may manipulate the data. The ADT is implementation
independent and enables the programmer to focus on idealized models of the data and its operations.
在项目设计中数据抽象是一个中心概念。定义了抽象域和数据结构,随着一个收集的操作,获取评价。抽象,叫了一个抽象数据类型(ADT),创建一个用户定义的数据类型,其业务指定客户如何操纵数据。ADT的是独立和可以让程序员,专注于理想化模型的数据和营运。
Examples 实例
An accounting program for a small business maintains inventory information. Each item in the inventory is represented by a data record that includes the item's identification number, the current stock level, pricing information, and reordering information.
一个会计程序为小企业维持库存的信息。库存的每一个项目是为一个数据记录,包括物品的数量,当前鉴定法进行库存水平、定价信息,并重新排序信息。
A set of inventory handling operations updates the different information fields and initiates a reordering of stock when inventory levels fall below a certain threshold.
一套库存的处理操作更新不同信息领域,重新启动一个股票当库存水平低于某一阈值。 The data abstraction describes an item as a record containing a series of information fields and operations that would be used by a company manager for inventory maintenance.
一个项目的数据抽象描述作为一个记录包含一系列的信息领域和操作,公司将使用存货维护经理。
Operations might include changing the Stock on Hand value when units of the item are sold,changing the Unit Price when a new pricing policy is used,and initiating a reorder when the stock level falls below the reorder level.
操作可能包括改变单位现货价值的物品被卖掉,改变单价的时候,一个新的价格政策,倡导使用重当库存水平低于重水平。
A gaming program involves tossing a set of dice. In the design, the dice are described as an ADT whose data includes the number of dice that are tossed,the sum of the dice on the last toss,and a list that identifies the value of each die in the last toss. The operations include tossing the dice, returning the sum of the dice on a toss,and printing the value of each individual die in the list.
一个游戏程序包括抛套骰子。在设计中被描述为一个ADT骰子,其数据的数量,包括投骰,增资额的骰子投掷,在最后一个列表识别每个骰子的数字在过去的搅拌。作业包括抛骰子,返回的金额在掷骰子,印刷它们各自的价值个人死在名单上。
Most abstract data types have an initializer operation that assigns initial values to the data. In the C++ language environment,the initializer is called a constructor.
最抽象数据类型有一个初始化器指定初始值的操作数据。在c++语言的环境中,被称为构造函数初始化器。 Stacks 堆栈
A stack is one of the most frequently used and most important data structures. Applications of stacks are vast. For instance, syntax recognition in a compiler is stack based, as is expression evaluation'-23. At a lower level, stacks are used to pass parameters to functions and to make the actual function call to and return from a function.
堆栈是其中一个最常用、最重要的数据结构。墙是巨大的应用。例如,在一个编译器语法识别的基础,是表达叠加评价。在一个较低的水平,墙是用来传递函数参数,使实际的函数调用一个函数和回归。
A stack is a list of items that are accessible at only one end of the list1-3-1. Items are added or deleted from the list only at the top of the stack. Food trays in a dining hall or a pile of boxes are good models for a stack.
堆栈是一种的商品清单只有一头可随时这份名单。项目是添加或删除从名单上只有在顶部的叠加。餐盘在饭厅或一堆箱子是理想模型堆栈。
A stack structure features operations that add and delete items. A Push operation adds an item to the top of the stack. The operation of removing an element from the stack is said to pop the stack. Figure 10-1 illustrates a sequence of Push and Pop operations. The last item added to the stack is the first one removed. For this reason,a stack is said to have LIFO (last-in/first-out)ordering.
堆栈结构特点,添加和删除操作物品。推操作,增加了一个项目的最上方。移除一个元素的操作堆栈是说从流行栈中。图10 - 1说明了一个序列的推进和流行歌剧其?。最后一项加到堆栈是第一个移除。因为这个原因,据说是一堆后进先出法)/(选择使用的排序。
The abstract concept of a stack allows for an indefinitely large list. Logically,food trays could be stacked to the sky. In reality, food trays are stored in a rack. When the rack is full,you cannot add (Push) another item on the stack. The stack has reached the maximum number of elements it can handle. The situation gives meaning to a \cannot pick up a tray from an empty stack. A\remove (Pop) an element
栈”这个抽象的概念提供了一个大名单。无限期从逻辑上讲,餐盘可以堆叠到天空。事实上,餐盘被存储在一个架子上。当架子上已经满了,您无法增加(推)另一项叠加。这个堆栈所包含的元素数量达到最大可以承受。意义的场合给一个”堆栈饱”的状态。在另一个极端,你不能捡起一个托盘从一个空的叠加。一个”堆栈空”状态意味着你不能解除(Pop)的元素
The Stack Class in C++ 栈类c++中
The stack members include a list,an index or pointer to the top of the stackM and the set of stack operations. We use an array to hold the stack elements. As a result,the stack size may not exceed the number of elements in the array and the stack full condition is relevant. We will remove the restriction when we develop a Stack class with a linked list.
堆栈的成员包括一个表,索引或指向顶层堆栈和堆栈操作的集合。我们使用一个数组来持有堆栈的元素。作为结果,堆栈大小不得超过该数组的元素个数和堆栈全部以相关的。我们将移除约束我们开发一种叠加一个链表的课。
The declaration of a stack object includes the stack size that defines the maximum number of elements in the list. The size has a default value MaxStackSize = 50. The list (stacklist) ,the maximum number of elements in the stack (size) and the index (top) are private members. The operations are public.
声明一个栈对象包括堆栈大小,定义了最多的元素在名单上。尺寸是一个默认值maxstacksize = 50。列表(栈列表),最多的元素在堆栈(大小号)和指数(top)是私人的成员。公开的操作。
Initially the stack is empty and top= — 1. Items enter the array (Push) in increasing order of the indices (top = 0,1,2) and come off the stack (Pop) in decreasing order of the indices (top =2,1,0).
最初的堆栈是空的,顶部=-1。物品进入阵列(推)指数增长的订单(= 0、1、2)和脱落堆栈(Pop)降序排列的指标(= 2、1、0)。
相关推荐：
loading