【文章內(nèi)容簡介】 A， R） 通常固有狀態(tài)將會通過存儲好的學(xué)習(xí)算法計算并用狀態(tài)表和行為數(shù)據(jù)的方式編碼。通過遵循調(diào)整當(dāng)前狀態(tài)的強(qiáng)化命令，升級功能計算當(dāng)前固有狀態(tài)、輸入、要執(zhí)中英文資料 行的行 為和新狀態(tài)的強(qiáng)化，并且根據(jù)存儲起來的固有狀態(tài)信息，評價功能 V計算固有狀態(tài)并輸入一個動作。在不同的強(qiáng)化學(xué)習(xí)中， U和 V有一些不同點(diǎn)。 Q 學(xué)習(xí)僅僅是一個典型的強(qiáng)化學(xué)習(xí)，在這種模式中，訓(xùn)練樣本并不跟 S,A相似，相似的是給機(jī)器人行為的獎勵。所以很難通過學(xué)習(xí)得到最好的∏ :S→ A 策略。這個問題可以通過學(xué)習(xí)定義在狀態(tài)和行為之上的評價功能來解決。通過學(xué)習(xí)評價功能，一個智能體獲得最基本的策略。一個很明顯的好的評價功能是 V*，當(dāng) V*（ S1） V*（ S2）時，智能體將會希望進(jìn)入 S1，因?yàn)槟菢訒玫揭粋€大一些的獎賞。當(dāng)然智能體會 選擇行為，而不會選擇狀態(tài)。 8 研究機(jī)器人杯的意義： 通過仔細(xì)考慮，我們可以提出更多的內(nèi)容和難題，我們看起來有理由拒絕“人類和機(jī)器之間的比賽”的想象，因?yàn)樵诋?dāng)前要達(dá)到這樣的目標(biāo)是不可能的。 但是回望歷史，過去有太多的我們先人無法想象的科學(xué)成就不是嗎？ 50 年內(nèi)人們將會對科學(xué)發(fā)展產(chǎn)生新的認(rèn)識。從第一架萊特兄弟的飛機(jī)試飛成功到阿波羅號飛船登陸月球，大約僅有半個世紀(jì)，同樣從第一臺電腦的誕生到“深藍(lán)”電腦打敗人類天才也是半個世紀(jì)?，F(xiàn)在我們應(yīng)該明白，在 50 年后，我們在“人類和機(jī)器之間的比賽”面前將不會說不。 我們現(xiàn)在需要 的是創(chuàng)新與積極參與，我們應(yīng)該做的是盡最大努力去促進(jìn)這個過程。 很明顯我們應(yīng)更加注重創(chuàng)新，機(jī)器人杯包含人工生命、能源、動力、材料等學(xué)科的發(fā)展，同樣它也孕育著機(jī)械、電子、控制、信息和電腦等與機(jī)器人相關(guān)的學(xué)科的重大突破。 機(jī)器人杯雖然只是幾個球員的高科技球賽，但它卻向我們展示了令人費(fèi)解的場景，比如機(jī)器人撞墻，兩個機(jī)器人相互糾纏，并且有些機(jī)器人顯得很迷亂，并不關(guān)心球在哪里。人們不理解為什么機(jī)器人的智力還趕不上個孩子。 這就是說讓機(jī)器人擁有人類的智慧及思考能力哪怕是三歲小孩的能力并不是一件容易的事情，到 2050 年，科學(xué)家們想組建一支全自動的能夠在與人類冠軍隊的比賽中獲勝的機(jī)器人隊伍。這是一個偉大的目標(biāo)。 結(jié)語： 這個論文討論了一些 MAS 系統(tǒng)和機(jī)器人杯的主要技術(shù)。這個主要目標(biāo)是讓讀者中英文資料 知道更多的關(guān)于智能體系統(tǒng)的知識和引發(fā)智能體導(dǎo)向技術(shù)的快速成熟的原因。 程序設(shè)計的發(fā)展總共有四個階段：面向過程的編程，模塊化編程，面向?qū)ο蟮木幊蹋椭悄芫幊?。每個過程都是越來越抽象越來越模糊的建模工程，直到最終實(shí)現(xiàn)自動編程。因此智能化的編程是編程史上不可避免的一個過程，機(jī)器人杯是一個促進(jìn)機(jī)器人基礎(chǔ)技術(shù)研究的動力。 致謝： 本作品獲得了中國 博士基金組織的支持。 參考文獻(xiàn)： [1]Balch T, Mhybite (2020), Social Potentials for Scalable MultiRobot Formation. IEEE International Robotics and Automation (ICRA 2020):7380. [2]Magnus Boman(1999), Agent Programming in RoboCup39。99. AgentLink NewsLetter, (4), November 1999. [3]Burkhard H D,et al (2020),The Road to RoboCup 2050. IEEE Robotics amp。 Automation Magazine. Jun. 2020: 3138. [4]Cai Qingsheng, Zhang Bo (2020), An agent team reinforcement learning model and its application. (J). Journal of Computer Research and Development. 2020,37(9): 1087 1093. In China. [5]Cheng Xianyi (2020), Agent Computing. Haerbin(China): Hei Longjiang science and technology press. 2020. [6]Cheng Xianyi et al(2020),.Reinforcement Learning in Simulation RoboCup Soccer. Proceeding of 2020 International Conference on Machine Learning and Cyberics (ICML2020),in China, IEEE Catalog Number:04EX826. pp244248. [7]Fredrik Heintz (2020), RoboSoc a System for Developing RoboCup Agents for Educational Use. Master39。s thesis, Department of Computer and Information Science, university, March 2020. [8]Hiroaki Y et al (2020), A Distributed Control Scheme for Multiple Robotic Vehicles to Make Group Forma and Autonomous systems,2020, 125 –147. Silvia Coradeschi and Jacek Malec(1999), How to make achallenging AI course enjoyable using the RoboCup soccer simulation system. In RoboCup98:The Second 中英文資料 Robot World Cup Soccer Games and Conference, pages 120{124. Springer verlag, 1999. [9]Johan Kummeneje, David , and H_akan L. Younes (1999), UBU – an object oriented RoboCup Team. In Johan Kummeneje and Magnus Boman, editors, Int7 1999 Papers. 1999. [10]Johan Kummeneje (1999), Simulated Robotic Soccer and the Use of Sociology in Real Time Mission Critical Systems. In L. R. Welch and M. W. Masters,editors, Proceedings of RTMCS Workshop, IEEE, December 1999 中英文資料 RoboCup is a Stage which Impulse the Research of Basic Technology in Robot Cheng Xianyi amp。 Xia Deshen 1. Introduction RoboCup is an international joint project to promote Artificial Intelligence (AI), robotics, and related field. It is an attempt to foster AI and intelligent robotics research by providing a standard problem where wide range of technologies can be integrated and examined. RoboCup chose to use soccer game as a central topic of research, aiming at innovations to be applied for socially significant problems and industries. The ultimate goal of the RoboCup project is by 2050, develop a team of fully autonomous humanoid robots that can win against the human world champion team in soccer (Fig 1). In order for a robot team to actually perform a soccer game, various technologies must be incorporated including: design principles of autonomous agents, multiagent 中英文資料 collaboration, strategy acquisition, realtime reasoning, robotics, and sensorfusion. RoboCup is a task for a team of multiple fastmoving robots under a dynamic environment. RoboCup also offers a software platform for research on the software aspects of RoboCup (Burkhard02). One of the major applications of RoboCup technologies is a search and rescue in largescale disaster. RoboCup initiated RoboCup rescue project to specifically promote research in socially significant issues. In the next section we will introduce the origin, anization and leagues of RoboCup. Section 3 we will discuss the relative technology in RoboCup. 2. The Origin, Organization and Leagues of Robocup The concept of RoboCup was first introduced by professor of Alan Mackworth in 1993. The main goal of RoboCup is to propose a challenged research issue to develop robotic. 中英文資料 Following a twoyear feasibility study, in August 1995, an announcement was made on the introduction of the first international conferences and football games. Now RoboCup Soccer is divided into the following leagues: Simulation league(2D,3D), Smallsize robot league (f180), Middlesize robot league (f2020), Fourlegged robot league, Humanoid league. Figure 1. Soccer racing in the future 434 In July 1997, the first official conference and games were held in Japan. The annual events attracted many participants and spectators. RoboCup 2DSimulation League The RoboCup 2Dsimulation league uses a simulator called the Soccer Server to do the soccer simulation. The Soccer Server provides a standard platform for research into multiagent systems. The Soccer Server simulates the players, the ball and the field for a 2D soccer clients (11 for each team) connect to the server, e