Research on the hottest artificial intelligence an

2022-08-01
  • Detail

Research on artificial intelligence and its application in intelligent manufacturing

research background of intelligent manufacturing

intelligent manufacturing (IM) originates from the research on artificial intelligence (AI). AI is a branch of computer science. Its emergence and development are closely related to the development of van Neumann computers in the past half century. Reviewing the development history of AI, it can be divided into three stages: gestation (before 1956), formation (1956~1969) and development (since 1970). Since ancient times, people have been trying to use a variety of machines to replace part of human mental work in order to improve their ability to conquer nature. The relevant viewpoints and achievements of predecessors gave birth to AI discipline. In particular, in 1936, the British mathematician a. m. Turing put forward a mathematical model of an ideal computer Turing machine; In 1950, he asked, "can computers think?" This article discusses that machines can think, so Turing is honored as the father of artificial intelligence. In 1956, at Dartmouth University in the United States, a young mathematician J. McCarthy and his three friends M. Minsky, n. lochester and C. Shannon jointly initiated the Summer Symposium, inviting scholars and scientific researchers from IBM, MIT, Rand and CMU. At the symposium, the term AI was officially adopted to represent the research direction of machine intelligence for the first time. This important meeting with historical significance marks the birth of a new discipline of AI. In the following ten years, AI has made many remarkable achievements in machine learning, theorem proving, pattern recognition, problem solving, expert system and AI language lisp. Since 1970, many countries have vigorously carried out AI research, and a large number of research results have emerged. AI has entered the development stage since then

The development of AI has had a far-reaching impact on human beings and the future. These impacts involve human economic interests, social roles and cultural life. Some people regard it, together with space technology and atomic energy technology, as the three major scientific and technological achievements of the 20th century; Some people also call it another revolution after the three industrial revolutions, and say that the first three revolutions have mainly extended the function of human hands and liberated human beings from heavy physical labor, while AI has extended the function of human brain and realized the automation of mental labor. At present, AI research is mostly carried out in combination with specific fields. Its research and application fields mainly include the following aspects: problem solving, logical reasoning and theorem proving, natural language understanding, automatic programming, machine learning, expert system, knowledge engineering, fuzzy logic, neural network, genetic algorithm, rough set theory, robotics, pattern recognition, machine vision, intelligent control, intelligent retrieval Intelligent dispatching and command, intelligent agent, system and language tools, etc

since the 1980s, the modern scientific and technological revolution has promoted the human society to enter the information society from the industrial society, making the modern manufacturing system change from the original energy driven type to the information driven type, which requires that the manufacturing system should not only be flexible, but also show intelligence, otherwise it is difficult to deal with such a large number of complex information; Moreover, the requirements of multi variety, variable batch, flexible production, rapidly changing market demand and the complex environment of fierce competition also require the manufacturing system to be more flexible, agile and intelligent. In this context, people began to try to study the application of AI in manufacturing industry, and achieved a number of fruitful results. In the late 1980s, an intelligent manufacturing technology and intelligent manufacturing system developed by integrating manufacturing automation, AI, computer science and other high and new technologies came to the fore. Subsequently, intelligent manufacturing, as a new discipline, received more and more attention. It is widely recognized that the theoretical research and application development of intelligent manufacturing are of great significance for improving product quality, production efficiency and cost reduction, improving the ability and speed of the national manufacturing industry to respond to market changes, and improving the national economic strength and living standards. Therefore, governments of all countries have included intelligent manufacturing in their national development plans and vigorously promoted its implementation. In the late 1980s, China also included "intelligent simulation" in the main subject of the national science and technology development plan, and has made a number of achievements in expert system, pattern recognition, robot and Chinese machine understanding; Recently, the Ministry of science and technology of the people's Republic of China formally proposed the "industrial intelligence project", which is an important part of the innovation capacity building in the technological innovation plan. Intelligent manufacturing is an important part of the project

overview of intelligent manufacturing

conceptually, "manufacturing" in intelligent manufacturing is the concept of "large-scale manufacturing". It not only refers to the traditional meaning of processing and technology, but also includes activities at all stages of the product life cycle, including design, organization, supply, sales, scrap and recycling. In terms of connotation, intelligent manufacturing mainly includes intelligent manufacturing technology and intelligent manufacturing system. Intelligent manufacturing technology is a comprehensive technology formed by the mutual penetration and interweaving of manufacturing technology, automation technology, system engineering and artificial intelligence. Intelligent manufacturing system is an environment for integrated application of intelligent technology 5. The host adopts full plastic spraying shell technology, and it is also the carrier of intelligent manufacturing mode. In 1990, the Japanese Ministry of trade and industry first put forward the concept of intelligent manufacturing system. The intelligent manufacturing system is a man-machine integration system composed of intelligent machines and human experts. It highlights that in all links of manufacturing, in a highly flexible and integrated way, with the help of the intelligent activities of human experts simulated by computers, it can analyze, judge, reason, conceive and make decisions, replace or extend part of human mental work in the manufacturing environment, and collect, store, improve, share Inherit and develop the manufacturing intelligence of human experts. Because this manufacturing mode highlights the value position of knowledge in manufacturing activities, and knowledge economy is the main economic form after industrial economy, intelligent manufacturing has become an advanced manufacturing production mode that affects the future economic development process

research status and development trends of artificial intelligence

modern computers have strong arithmetic and logic operation functions. Especially in long-time operations, the operation speed is as high as hundreds of billions of times/second, and the accuracy and reliability of the results are beyond the reach of human beings. Therefore, in the early AI research, the use of symbols, rule expression and logical reasoning has become the mainstream, These studies are also known as the "semiotic" School of AI. Although the traditional AI based on van Neumann's computer has made some achievements in expert systems, game games, theorem proving, etc., so far, the intelligence embodied by AI is still far away from the human brain as a whole, especially in the visual thinking ability of computer. Therefore, the study of neural network, which simulates the image thinking of human brain and the structure of nervous system, came into being. In 1943, psychologist mccalloch cooperated with mathematician Pitts to put forward the first mathematical calculation model of neural networks (NN) - MP model, thus creating a new era of neural network theory research. For more than 50 years, although the research on neural network once fell into a trough, since J. Hopfield put forward the Hopfield neural network model in 1982, which successfully solved the traveling salesman problem, its research has no need to enter the climax stage again, and is developing towards self-organization, self-adaptive, self-learning and so on. The study of neural network is also known as the "connectionist" School of AI. In 1975, john H. Holland proposed an optimization algorithm genetic algorithm (GA) based on the model of biological evolution, which is an adaptive heuristic search technology in a global sense. It evolves a better solution that meets the given accuracy after heredity and mutation according to the natural selection law of the survival of the fittest in nature. The research on genetic algorithms, evolutionary programming and evolutionary strategies is also called "behaviorism" School of AI

symbolism only grasps the characteristics of abstract thinking of human brain; Connectionism only imitates people's thinking in images; Behaviorism? Considering the characteristics of human intelligent behavior and evolution process, they all have obvious limitations, and their application levels and occasions are relatively different. In order to make the machine reach the multi-functional, multi-level, multi-faceted and all-round advanced intelligent behavior of human brain, we must study intelligence from the multi-dimensional and global point of view in order to overcome the above limitations. Since the 1990s, people have begun to organically combine the various schools of AI, believing that this is an effective way to give full play to the advantages of the various schools of AI and make up for their shortcomings. The following is a brief introduction to the research trends of AI in realizing intelligent theory and methods in recent years

the combination of fuzzy logic and neural networks

in 1967, Professor L. A. Zadeh of the University of California published a pioneering paper entitled "fuzzy sets", and then developed fuzzy mathematics based on fuzzy sets into a special discipline. Neural network and fuzzy logic (FL) have complementary advantages and disadvantages, as shown in Table 2. In 1974, American professors C. T. Lin and C. S. g. Lee first introduced fuzzy thought into neural network. In 1987, B. Kosko took the lead in systematically combining fuzzy logic with neural network. Fuzzy neural network (FNN) is the product of the organic combination of the two. It combines the advantages of the two, integrating learning, association, recognition, adaptation and fuzzy information processing. So far, there are many ways to fuse neural network and fuzzy logic to form FNN, but there are two kinds of FNN systems that truly embody the advantages of both: first, the system is supported by fuzzy logic, and the fuzzy logic rules are realized by neural network; Second, the system uses neural network structure, and the learning process uses fuzzy reasoning. In the first kind of FNN, the ability of neural network to learn mapping rules from input/output samples is used to establish the rules required by fuzzy logic, so as to solve the "bottleneck" problem of fuzzy rule extraction in fuzzy logic system. In the second type of FNN, the commonly used value of neural network is changed into membership input, or simple max min operation is used to replace the complex product sum operation of neural network, and fuzzy rules are used to guide the learning of neural network, which speeds up the learning process of neural network and makes the logical analysis of the system more explicit

combination of genetic algorithm and neural network

genetic algorithm formulates guided optimization strategy or search algorithm according to the law of survival of the fittest in biological evolution theory and the basic behavior characteristics of life, such as self-learning, self-organization and self adaptation. Its characteristic is that the population most suitable for a specific objective function is reorganized in a random form to produce a new generation, and there are mutation steps such as selection, reorganization and replacement in the evolution process, To search for better sample space points. In recent years, combining genetic algorithm with the learning of connection weights of neural networks and the optimization of neurons and network structures, the study of robust evolutionary neural networks has become a new research hotspot in the field of neural networks, and has made many valuable conclusions and achievements, which brings a promising prospect for the wide and in-depth application in engineering

the combination of genetic algorithm and fuzzy logic

the combination of genetic algorithm and fuzzy logic is a new field that organically combines knowledge acquisition and knowledge representation. The current research focus is: using genetic algorithm to improve fuzzy logic system, such as genetic algorithm

Copyright © 2011 JIN SHI