Model Decomposition and Simulation

IEEE Transactions on Systems, Man, and Cybernetics, 1992

A qualitative model of an expert's mental model of a complex system (advanced nuclear power plant) was developed from the qualitative physics of confluences. This model was implemented as a qualitative simulation using an object-oriented extension to Common Lisp (Flavors). An existing method for dynamic constraint satisfaction was found to be inadequate for complex systems. Invisible connections for flow compatibility, control connections, iterative propagation, and embedded propagation were among the new features provided for derivation of causal ordering. Deterministic output was guaranteed through stochastic state transition. A parametric Monte Carlo simulation study was performed using a fictitious loop fragment, and changes were observed in flow rate change through a pump. State transition models provided excellent fits to the simulation data. Analysis of the state models showed that all conditions converged to steady state. Strictly forward (with the now) propagation facilitated consistency within intermediate pre-equilibrium states and convergence as compared to forward propagation with limited backward propagation. Uncertainty bias inhibited propagation of premature incorrect values. The psychological plausibility of qualitative simulation models was evaluated. A further extension of mythical causality is suggested, for which constraint propagation executes on multiple levels of aggregation.

2005

This paper presents an approach to generate structured explanations of system behaviour based on qualitative simulations. This has been implemented in WiziGarp, a domain-independent interactive learning environment. The main issue addressed here is how to manage the complexity of a simulation in order to generate adequate explanations. These are presented to the user in the form of different kinds of diagrams, accompanied by explantory dialogue.

2010

This paper presents a divide-and-conquer approach that aims at making QSIM simulation tractable. We consider dynamical systems the structure of which can be represented by compartments. The system model is decomposed into sub-models tightly connected through shared variables on the basis of the analysis of the causality relations intrinsically captured by the compartmental model structure. The sub-models are separately simulated but their behaviors are constrained by the information on the shared variables generated from the simulation. The partition of the complete model into smaller ones prevents the construction of temporal correlations between variables in different sub-models, and thus the generation of a complete temporal ordering of all unrelated events that is one of the major causes of intractable branching in qualitative simulation. The strategy we propose is discussed through a case study in the eld of Plant Pathology, namely the germination process of Plasmopara viticola...

Simulation Modelling Practice and Theory, 2006

Discrete event simulation is an important system analysis technique. But for today's demand for speed, the time to complete a simulation study is considered to be long, even with current developments in hardware and simulation software. In this scenario, simplification methods for simulation models could play a key role. This paper proposes a technique for reducing the complexity of a discrete event simulation model at the conceptual phase of simulation modeling that can be fully automatized through a computer program. We applied this technique on some problems which demonstrate the feasibility of this approach. (L. Chwif). Simulation Modelling Practice and Theory 14 (2006) 930-944 www.elsevier.com/locate/simpat

Qualitative simulation is a well-known reasoning technique that involves the use of simulation technologies. Reasoning is made to determine qualitative values and change directions of system variables, and it is done for each time point and time interval following the time point. Qualitative variables possess continuous qualitative value sets that are discretized by landmark points. Qualitative simulation uses qualitative time representation and its quantitative value is of no interest. The main purpose of this study was to develop a technique to determine time steps for a quantitative simulation under guidance of qualitative information. The proposed technique determined time advances using qualitative and quantitative information together to obtain a robust time step as wide as possible for simulation time advances. For this purpose, sign algebraic properties and derivation roots of quantitative equations and qualitative variable values with their change directions were used to compute time advances. In the approach, qualitative simulation determined landmark points to be advanced, and quantitative simulation calculated the duration required. Using the proposed algorithm, the simulation is advanced instead of iterating simulation time for a predefined time step and checking whether or not there is any activity in the interval, directly to the time points that are qualitatively different.

Artificial Intelligence, 1986

Qualitative simulation is a key inference process in qualitative causal reasoning. However, the precise meaning of the different proposals and their relation with differential equations is often unclear. In this paper, we present a precise definition of qualitative structure and behavior descriptions as abstractions of differential equations and continuously differentiable functions. We present a new algorithm for qualitative simulation that generalizes the best features of existing algorithms, and allows direct comparisons among alternate approaches. Starting with a set of constraints abstracted from a differential equation, we prove that the qsim algorithm is guaranteed to produce a qualitative behavior corresponding to any solution to the original equation. We also show that any qualitative simulation algorithm will sometimes produce spurious qualitative behaviors: ones which do not correspond to any mechanism satisfying the given constraints. These observations suggest specific types of care that must be taken in designing applications of qualitative causal reasoning systems, and in constructing and validating a knowledge base of mechanism descriptions.

1997

Abstract One of the major factors hindering the use of qualitative simulation techniques to reason about the behavior of complex dynamical systems is intractable branching due to a phenomenon called chatter. This paper presents two general abstraction techniques that solve the problem of chatter. Eliminating the problem of chatter signi cantly extends the range of models that can be tractably simulated using qualitative simulation.

Artificial Intelligence, 1993

Qualitative reasoning about physical systems has become one of the most active and productive areas in AI in recent years. While there are many different kinds of qualitative reasoning, the central role is played by qualitative simulation: prediction of the possible behaviors consistent with incomplete knowledge of the structure of physical system. In the retrospective [8] on my 1984 paper, "Commonsense reasoning about causality: deriving behavior from structure", I describe the framework for qualitative reasoning that has motivated this work, and the applications that have come out of that framework. That paper [5 ] includes the conjecture that the structural and behavioral representations for qualitative simulation could be rigorously shown to be abstractions of ordinary differential equations and their solutions. My 1986 paper, "Qualitative simulation", established that conjecture and legitimized the term qualitative differential equation or QDE. It also presented the clear and efficient QSIM algorithm. In this retrospective, I describe aspects of the body of work on qualitative simulation that has developed from there.

2005

This paper presents and discusses work on the automated generation of qualitative (diagnostic) models from simulation models that have been developed for (control) engineering purposes. This work is motivated by an attempt to build model-based tools that support a closer integration of diagnostic considerations in early design phases of on-board systems for vehicles and based on the insight that such an attempt has to limit the required modeling efforts. We present the mathematical foundations and the implementation of the abstraction process and discuss the various difficulties and problems encountered when we applied the software to real automotive subsystems. These difficulties include complexity and methodological issues, and what should be, but has not been, a major concern of research on qualitative reasoning: How to obtain adequate qualitative domains.

SIMULATION, 2004

A crucial decision within component-based modeling and simulation is the choice of the criterion or criteria to be used in decomposing the system. Many experiences with combat simulation systems indicate, first, that the optimal choice depends on the driving forces behind the decomposition or the goals that should be achieved by using components instead of monolithic systems. Second, the common criteria for decomposition in software design, information hiding, and “object picking”are sometimes inappropriate for model decomposition because they neglect the importance of so-called hidden assumptions of model abstraction and idealization, which are paramount to the modeling of every complex system.

1991

Abstract: One main problem in qualitative simulation is that it often produces too detailed qualitative descriptions of a system's possible behaviors, or produce a lot of behaviors that differ very slightly. This paper addresses the problem of summarizing the result of a qualitative simulation to make it more perspicuous. Two causes of behavior proliferation have been identified and two algorithms to aggregate behaviors that do not differ significantly are presented.

International Joint Conference on Artificial Intelligence, 1985

Qualitative simulation is a key inference process in qualitative causal reasoning, In this paper, we present the QSIM algorithm, a new algorithm for qualitative simulation that generalizes the best features of rxisting algorithms, and allows direct comparisons among alternate approaches. QSIM is an efficient constraint-satisfaction algorithm that can follow either its standard semantics allowing the creation of new landmarks, or the {+, 0, -} semantics where 0 is the only landmark value, by changing a table of legal state-transitions. We argue that the QSIM semantics make more appropriate qualitative distinctions since the { + ,0,-} semantics can collapse the distinction among increasing, stable, or decreasing oscillation. We also show that (a) qualitative simulation algorithms can be proved to produce every actual behavior of the mechanism being modeled, but (b) existing qualitative simulation algorithms, because of their local points of view, can predict spurious behaviors not produced by any mechanism satisfying the structural description. These observations suggest specific types of care that must be taken in designing applications of qualitative causal reasoning systems, and in constructing and validating a knowledge base of mechanism descriptions.

2000

Qualitative simulations can be seen as knowledge models that capture insights about system behaviour that should be acquired by learners. A problem that learners encounter when interacting with qualitative simulations is the overwhelming amount of knowledge detail represented in such models. As a result, the discovery space grows too large, which hampers the knowledge construction process of the learner. In

1988

Abstract Qualitative simulation faces an intrinsic problem of scale: the number of limit hypotheses grows exponentially with the number of parameters approaching limits. We present a method called Time-Scale Abstraction for structuring a complex system as a hierarchy of smaller, interacting equilibrium mechanisms. Within this hierarchy, a given mechanism views a slower one as being constant, and a faster one as being instantaneous.

Artificial Intelligence, 1993

Qualitative reasoning about physical systems has become one of the most productive areas in AI in recent years, due in part to the 1984 special issue of Artificial Intelligence on that topic. My contribution to that issue was a paper entitled "Commonsense reasoning about causality: deriving behavior from structure" [9]. From my perspective, that paper laid out a research program that has continued to be productive to this day, and promises to continue well into the future. After establishing a framework for qualitative reasoning, the primary technical contribution of the paper was a simple, clear representation for qualitative structure and behavior, abstracted from ordinary differential equations. My subsequent Artificial Intelligence paper, "Qualitative simulation" [ 10 ], made that abstraction relation precise, presented the vastly improved QSIM algorithm for qualitative simulation, and used the abstraction relation to prove the soundness and incompleteness of QSIM. I will discuss developments in qualitative simulation in my retrospective on that paper [12], and concentrate here on the larger issue of reasoning with qualitative models.