Constraints provide a useful mechanism for maintaining relations in user interface toolkits. Garnet is a widely-used user interface toolkit with considerable functionality, based on one-way, required constraints. Multi-Garnet extends Garnet by adding support for multi-way constraints and constraint hierarchies with both required and preferential constraints. This document contains three chapters describing Multi-Garnet: ffl Chapter 1 presents a high-level overview of Multi-Garnet. To motivate the development of Multi-Garnet, we examine the Garnet constraint system, present some realistic user interface problems that are difficult to handle in Garnet, and demonstrate how Multi-Garnet addresses these problems. We provide details on how Multi-Garnet supports some of the features of Garnet, including constraints with pointer variables, and inheritance of constraints. ffl Chapter 2 contains a reference manual for the current version of Multi-Garnet (version 2.1). This includes information ...

We present Pika, an implemented self-explanatory simulator that is more than 5000 times faster than SimGen Mk2 [ Forbus and Falkenhainer, 1992 ] , the previous state of the art. Like SimGen, Pika automatically prepares and runs a numeric simulation of a physical device specified as a particular instantiation of a general domain theory, and it is capable of explaining its reasoning and the simulated behavior. Unlike SimGen, Pika's modeling language allows arbitrary algebraic and differential equations with no prespecified causal direction; Pika infers the appropriate causality and solves the equations as necessary to prepare for numeric integration. Introduction Science and engineering have used numeric simulation productively for years. Simulation programs, however, have been laboriously hand-crafted, intricate, and difficult to understand and change. There has been much recent work on automating their construction (e.g. [ Yang, 1992, Rosenberg and Karnopp, 1983, Abelson...

This paper proposes the development of a semi-qualitative simulation framework for virtual environments. Its purpose is to simulate the behavior of complex devices and their interaction with virtual humans. We present an object-oriented semi-qualitative language to model such system. While our use of hierarchical finite state machines allows for the development of complex components, we retain the expressiveness of the compositional modeling used in qualitative physics. Using high level primitives, a physics-based model can directly animate 3D geometry. Participating agents can operate, assemble or disassemble a device. They can also access qualitative models and simulation histories to reason about its behavior. A potential application of this framework is in maintenance simulation. Other applications include virtual laboratories and virtual prototyping. Contents 1 Introduction 1 1.1 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.1 Modeling . . . . . . . . ...

3D virtual environments are being used in an increasing array of applications ranging from training systems, to video games and social communities. The steady growth of computing power and rendering capabilities makes them even more compelling as eachday passes. These visually richworlds are inhabited byintelligent agents and contain physicsbased simulated artifacts ranging from a light switch to a complete aircraft. However, a whole category of applications, such as maintenance simulations and virtual construction sets, based on the ability to alter on-the-#y the structure of these devices has yet to be developed. The main reason is that for performance purposes physics-based models must be pre-compiled at the time the application is developed, and thus they cannot be altered at run time. We present the development of an interactive semi-qualitative simulation framework to support assembly-centric virtual environments. The framework's purpose is to simulate the physics-based behavior of complex devices and their interaction with virtual humans. In particular, we use automated model building methods to enable users and participating agents to alter the structure of a device while its behavior is kept physically consistent without suspending the simulation. For example disconnecting a pipe in a virtual hydraulic system may cause the #uid it carries to spill in the environment. The framework v consists of a simulation engine and its object-oriented modeling language. The language captures the hybrid behavior of complex devices with hierarchical finite state machines. It encodes their continuous operation modes as di#erential algebraic equation systems. It ...