• Title/Summary/Keyword: Temporal abstraction

Search Result 18, Processing Time 0.027 seconds

Efficient Approximation of State Space for Reinforcement Learning Using Complex Network Models (복잡계망 모델을 사용한 강화 학습 상태 공간의 효율적인 근사)

  • Yi, Seung-Joon;Eom, Jae-Hong;Zhang, Byoung-Tak
    • Journal of KIISE:Software and Applications
    • /
    • v.36 no.6
    • /
    • pp.479-490
    • /
    • 2009
  • A number of temporal abstraction approaches have been suggested so far to handle the high computational complexity of Markov decision problems (MDPs). Although the structure of temporal abstraction can significantly affect the efficiency of solving the MDP, to our knowledge none of current temporal abstraction approaches explicitly consider the relationship between topology and efficiency. In this paper, we first show that a topological measurement from complex network literature, mean geodesic distance, can reflect the efficiency of solving MDP. Based on this, we build an incremental method to systematically build temporal abstractions using a network model that guarantees a small mean geodesic distance. We test our algorithm on a realistic 3D game environment, and experimental results show that our model has subpolynomial growth of mean geodesic distance according to problem size, which enables efficient solving of resulting MDP.

An Abstraction Method for State Minimization based on Syntactic and Semantic Patterns in the Execution Space of Real-Time Systems (실시간 시스템의 실행 공간상에서 구문 및 의미패턴에 기반한 상태 최소화를 위한 추상화 방법)

  • 박지연;조기환;이문근
    • Journal of KIISE:Software and Applications
    • /
    • v.30 no.1_2
    • /
    • pp.103-116
    • /
    • 2003
  • States explosion due to composition of spaces of data, temporal, and locational values is one of the well-known critical problems which cause difficulty in understanding and analysing real-time systems specified with state-based formal methods. In order to overcome this problem, this paper presents an abstraction method for state minimization based on an abstraction in system specification and an abstraction in system execution. The first is named the syntactic in system specification and an abstraction in system execution. The first is named the syntactic abstraction, through which the patterns of the unconditionally internalized computation and the repetition and selection structures are abstracted. The latter is named the semantic abstraction, through which the patterns of the execution space represented with data. Through the abstractions, the components of a system in specification and execution model is hierarchically organized. The system can be analyzed briefly in the upper level in an skeleton manner with low complexity. The system, however, can be abstraction method for the state minimization and the decrease in analysis complexity through the abstraction with examples.

Mitigating the State Explosion Problem using Relay Model Checking (릴레이 모델 체킹을 이용한 상태 폭발 문제 해결)

  • 이태훈;권기현
    • Journal of KIISE:Software and Applications
    • /
    • v.31 no.11
    • /
    • pp.1560-1567
    • /
    • 2004
  • In temporal logic model checking, the number of states is exponentially increased by the size of a model. This is called the state explosion problem. Abstraction, partial order, symmetric, etc. are widely used to avoid the problem. They reduce a number of states by exploiting structural information in a model. Instead, this paper proposes the relay model checking that decomposes a temporal formula to be verified into several sub-formulas and then model checking them one by one. As a result, we solve complex games that can't handle with previous techniques.

Overall Conversion Efficiency for Dimethylsulfide to Sulfur Dioxide in the Marine Boundary Layer-An Overview

  • Shon, Zang-Ho
    • Journal of Korean Society for Atmospheric Environment
    • /
    • v.18 no.E2
    • /
    • pp.107-120
    • /
    • 2002
  • Dimethyl sulfide (DMS) is the major sulfur gas released from the ocean. The atmospheric DMS released from the ocean is oxidized mainly by hydroxyl (OH) radical during the day and nitrate (NO$_3$) radical at night to form sulfur dioxide (SO$_2$) as well as other stable products. The oxidation mechanism of DMS via OH has been known to proceed by two channels; abstraction and addition channels. The major intermediate product of the addition channel has been known to be dimethylsulfoxide (DMSO) based on laboratory chamber studies and field experiments. However, a branching ratio for DMSO formation is still uncertain. The reaction of DMSO with OH ultimately produces SO$_2$and dimethylsulfone. The major product of the abstraction channel has known to be SO$_2$from laboratory chamber studies. But overall conversion efficiency for DMS to SO$_2$from DMS oxidation is still inconsistent in the literature. Based on laboratory and field studies, the conversion efficiency from the abstraction channel is likely to be greater than 0.5, while that from the addition channel is likely to be greater than 0.6. Overall conversion efficiency from DMS to SO$_2$might be greater than 0.5 based on the above two values in the remote marine boundary layer (MBL). This high efficiency in the remote MBL is supported by strong coupling between DMS and SO$_2$measurements with high temporal resolution.

Disambiguiation of Qualitative Reasoning with Quantitative Knowledge (정성추론에서의 모호성제거를 위한 양적지식의 활용)

  • Yoon, Wan-Chul
    • Journal of Korean Institute of Industrial Engineers
    • /
    • v.18 no.1
    • /
    • pp.81-89
    • /
    • 1992
  • After much research on qualitative reasoning, the problem of ambiguities still hampers the practicality of this important AI tool. In this paper, the sources of ambiguities are examined in depth with a systems engineering point of view and possible directions to disambiguation are suggested. This includes some modeling strategies and an architecture of temporal inference for building unambiguous qualitative models of practical complexity. It is argued that knowledge of multiple levels in abstraction hierarchy must be reflected in the modeling to resolve ambiguities by introducing the designer's decisions. The inference engine must be able to integrate two different types of temporal knowledge representation to determine the partial ordering of future events. As an independent quantity management system that supports the suggested modeling approach, LIQUIDS(Linear Quantity-Information Deriving System) is described. The inference scheme can be conjoined with ordinary rule-based reasoning systems and hence generalized into many different domains.

  • PDF

Formal Method for Specification and Verification of Behavioral Equivalences of Real-time Navigation and Transportation Systems Based on Abstraction (추상화에 기반을 둔 실시간 항법 및 배송 시스템의 명세 및 행위적 동일성 검증을 위한 정형 기법)

  • Lee, Moon-Kun;Choi, Jung-Rhan
    • The Journal of the Korea Contents Association
    • /
    • v.6 no.11
    • /
    • pp.202-216
    • /
    • 2006
  • A number of process algebras are not well suitable for real-time navigation/delivery systems due to the following reasons: 1) lack of representation of process distributivity over some geographical space and 2) the indistinction of representation of process mobility from process distributivity over the space. To make the process algebra suitable to the systems, it seems to be necessary to separate the space representation from the mobility representation. This paper presents a formal method for this purpose, namely, Calculus of Abstract Real-Time Distribution, Mobility, and Interaction (CARDMI). For analysis and verification of behavioral properties, CARDMI defines a set of the spatial, temporal and the interactive deduction rules and a set of equivalence relations. The rules and equivalences can be abstracted hierarchically due to the spatial abstraction, too. CARDMI can be applied to virtual navigation/delivery system for contents, too.

  • PDF

Reductions of State Space for Solving Games (게임 풀이를 위한 상태 공간 축소)

  • Lee, Tae-Hoon;Kwon, Gi-Hwon
    • Journal of Korea Game Society
    • /
    • v.4 no.1
    • /
    • pp.58-66
    • /
    • 2004
  • This paper uses counterexamples for solving reachability games. An objective. of the game we consider here is to find out a minimal path from an initial state to the goal state. We represent initial states and game rules as finite state model and the goal state as temporal logic formula. Then, model checking is used to determine whether the model satisfies the formula. In case the model does not satisfy the formula, model checking generates a counterexample that shows how to reach the goal state from an initial state. In this way, we solve many of small-sized Push Push games. However, we cannot handle larger-sized games due to the state explosion problem. To mitigate the problem, abstraction is used to reduce the state space to be che cked. As a result, unsolved games are solved with the abstraction technique we propose inthis paper.

  • PDF

On the Design of Distributed Time-Triggered Embedded Systems

  • Kopetz, Hermann
    • Journal of Computing Science and Engineering
    • /
    • v.2 no.4
    • /
    • pp.340-356
    • /
    • 2008
  • The cognitive constraints of the human mind must drive the decisions in architecture and methodology design in order that the systems we build are comprehensible. This paper presents a methodology for the design of time-triggered embedded systems that leads to understandable artifacts. We lift the design process to a higher level of abstractionto the level of computational components that interact solely by the exchange of messages. The time-triggered architecture makes it possible to specify the temporal properties of component interfaces precisely and provides temporally predictable message communication, such that the precise behavior of a large design can be studied in the early phases of a design on the basis of the component interface specifications. This paper shows how the cognitive simplification strategies of abstraction, partitioning and segmentation are supported by the time-triggered architecture and its associated design methodology to construct evolvable embedded systems that can be readily understood and modified.

A Formal Modeling of Managed Object Behaviour with Dynamic Temporal Properties (동적 시간지원 특성을 지원하는 망관리 객체의 정형적 모델링)

  • Choi, Eun-Bok;Lee, Hyung-Hyo;Noh, Bong-Nam
    • The Transactions of the Korea Information Processing Society
    • /
    • v.7 no.1
    • /
    • pp.166-180
    • /
    • 2000
  • Recommendations of ITU-T and ISO stipulate the managerial abstraction of static and dynamic characteristics of network elements, management functions as well as management communication protocol. The current recommendations provide the formal mechanism for the structural parts of managed objects such as managed object class and attributes. But the current description method does not provide the formal mechanism for the behavioral characteristics of managed objects in clear manner but in natural language form, the complete specification of managed objects is not fully described. Also, the behaviour of managed objects is affected by their temporal and active properties. While the temporal properties representing periodic or repetitive internals are to describe managed objects behaviour in rather strict way, it will be more powerful if more dynamic temporal properties determined by external conditions are added to managed objects. In this paper, we added dynamic features to scheduling managed objects, and described, in GDMO, scheduling managed objects that support dynamic features. We also described behaviour of managed objects in newly defined BDL that has dynamic temporal properties. This paper showed that dynamic temporal managed objects provide a systematic and formal method in agent management function model.

  • PDF

A Formal Modeling for Temporal and Active Properties of Managed Object Behavior (망관리 객체의 시간지원 능동 특성에 대한 전형적 모델링)

  • Choe, Eun-Bok;Lee, Hyeong-Ho;No, Bong-Nam
    • The Transactions of the Korea Information Processing Society
    • /
    • v.6 no.9
    • /
    • pp.2479-2492
    • /
    • 1999
  • Network management system(NMS) provides not only effective monitoring and controlling of network which consists of heterogeneous network elements but prompt response to users' need for high-level communication services. Recommendations of ITU-T and ISO stipulate the managerial abstraction of static and dynamic characteristics of network elements, management functions as well as management communication protocol. But the current description method does not provide the formal mechanism for the behavior characteristics of managed objects in clear manner but in natural language form, the complete specification of managed objects is not fully described. In this paper, we describe determinants for the behaviour of managed objects applicable to every managed object, and present a language for specifying behavioral aspects of managed objects based on their temporal and active properties.

  • PDF