“先期推理反应系统”之原始创新论文

程京德

这是本人 2002 年和 2004 年的两篇论文，开创了“先期推理反应系统”这一新研究方向，是笔者研究室在这一研究方向上所有工作的初始基础。现将两篇论文存放于此，一方面留作保存记录，另一方面提供给听课学生和有兴趣者获取。

顺便把论文概要翻译为中文如下：

ICSDS02 论文： 本文提出了一种新型的反应系统，称为“先期推理反应系统”，作为一类先期系统。本文介绍了开发先期推理反应系统的背景和动机，提出了系统的架构和逻辑基础，并展示了其可能的应用。

J. Cheng, “Anticipatory Reasoning-Reacting Systems,” Proceedings of the International Conference on Systems, Development and Self-organization, pp. 161-165, Beijing, China, November 2002, also in Proceedings of IPSJ SIGICS and IEICE SIG-AI JointMeeting, IPSJ SIG Technical Reports, Vol. 2003, No. 8, ICS-131, pp. 209-214, IEICE Technical Report, Vol. 102, No. 615, pp. 71-76, IPSJ and IEICE, January 2003.

Abstract：“This paper proposes a new type of reactive systems, named “Anticipatory Reasoning-Reacting Systems,” as a certain class of anticipatory systems.  The paper presents the background and motivation to develop anticipatory reasoning-reacting systems, proposes the architecture and logical basis of the systems, and shows their possible applications.”

PDF: ARRS_ICSDS02.pdf

CASYS03 论文： 一种名为“先期推理反应系统”的新型反应系统已经被提出作为一类先期系统。先期推理反应系统的一个关键功能是其先期推理的能力，它需要一个健全的逻辑基础为推理的有效性提供令人满意的标准。本文指出为什么经典数理逻辑、其各种古典保存扩张以及其各种非经典替代都不适合作为支撑先期推理的逻辑基础，说明时态相关逻辑是一个更有希望的候选；并简要介绍了名为“EnCal”的通用归结演算自动前向演绎系统，它可用作先期推理的推理引擎。本文还提出了一个新的和具有挑战性的研究问题（NRT问题）：我们如何在一个统一的逻辑系统中处理规范性、相关性和时间性这些模态概念，以表达、描述、推论和验证先期系统的先期行为？

J. Cheng, “Temporal Relevant Logic as the Logical Basis of Anticipatory Reasoning-Reacting Systems” (Invited Paper), in D. M. Dubois (Ed.), “Computing Anticipatory Systems: CASYS 2003 -Sixth International Conference, Liege, Belgium, 11-16 August 2003,” AIP Conference Proceedings, Vol. 718, pp. 362-375, American Institute of Physics, August 2004. (Best Paper Award awarded at the Sixth International Conference on Computing Anticipatory Systems)

Abstract：“A new type ofreactive systems, named ‘Anticipatory Reasoning-Reacting Systems’, has been proposed as a certain class of anticipatory systems.  A key function of an anticipatory reasoning-reacting system is its ability of anticipatory reasoning that requires a sound logic basis to provide a satisfactory criterion for validityof reasoning.  This paper points out why the classical mathematical logic, its various classical conservative extensions, and its various non-classical alternatives are not suitable logical basis for underlying anticipatory reasoning, shows that temporal relevant logicis a more hopeful candidate for the purpose, and gives a brief introduction to an automated forward deduction system for general-purpose entailment calculus, named ‘EnCal’, which can be used as a reasoning engine for anticipatory reasoning.  The paper also proposes a new and challenging research problem (NRT problem): how can we deal with those modal notions of Normativeness, Relevance, and Temporality within a unified logic system to underlie representing, specifying, reasoning about, and verifying anticipatory behavior of an anticipatory system?”

PDF: ARRS_TRL_CASYS03AIP718.pdf