Introduction of Simulation Learning and Optimization to Support Engineering Design

本研究的目的是将模拟、学习和优化相结合，形成一个自适应和交互建模系统(AIMS)，以支持不同阶段的设计和并行工程的多角度。在给定从模拟或实验中收集的实例数据库以及一组建模目标和设计目标的情况下，AIMS在模型形成和使用阶段执行多目标优化。通过不同的建模目标，如速度、精度和可理解性，AIMS可以生成分层模型，以满足设计不同阶段不断变化的需求。通过将不同的生命周期关注点视为相互竞争的设计目标，AIMS可以支持并行工程的多个视角的系统整合。这项研究是基于人工智能的技术与传统工程方法的一次重大而有意义的结合，以进一步推动设计自动化的基础研究。

The objective of this research is to integrate simulation, learning, and optimization to form an Adaptive and Interactive Modeling System (AIMS) to support multiple perspectives at different stages for design and concurrent engineering. Given a database of examples collected from simulations or experiments, and a set of modeling objectives and design objectives, AIMS performs multi-objective optimization at both model formation and utilization phases. By varying modeling objectives, such as speed, accuracy, and comprehensibility, AIMS can produce layered models to meet the changing needs at different stages of design. By treating different life-cycle concerns as competing design objectives, AIMS can support systematic incorporation of multiple perspectives for concurrent engineering. This research represents a significant and meaningful integration between artificial intelligence based techniques and conventional engineering methods to further advance the basic research in design automation.