Understanding Information Acquisition Decisions in Systems Design through Behavioral Experiments and Bayesian Analysis

点击翻译按钮获取中文摘要

The primary research objective in this project is to understand how individuals make information acquisition decisions in engineering systems design. Information acquisition is a key activity within systems engineering and design. It involves decisions such as whether or not to gain more information about a design concept, whether to execute a simulation or to run a physical experiment, and selecting from alternate ways to refine a behavioral model of a system. Information acquisition decisions have a significant effect on the quality of design outcomes and the resources needed to achieve the outcomes. While there has been significant progress in understanding how such decisions should ideally be made, there is a significant gap in knowledge about how humans actually make such decisions. This gap is a barrier to improving systems engineering and design practice. In this project, basic research towards addressing this gap will be carried out. Through a combination of theories from psychological and cognitive sciences, and empirical evidence from individual decisions within different design situations, the project will provide fundamental understanding of how humans make decisions in systems design, and result in explanatory models for how those decisions deviate from ideal behavior. On successful completion, the project will have three specific outcomes. First, a consistent analytical framework for describing strategies followed by humans in design-related information acquisition decisions and the effects of problem-specific and individual-specific influencing factors will be established. Second, the project will result in an experimental framework consisting of a set of behavioral experiments based on engineering design problems, instantiated as games and implemented in an online platform, for efficiently conducting behavioral experiments in the lab and in the field. Third, a reasoning framework will be established that probabilistically represents the state of knowledge about which descriptive models best represent individuals' design decisions, and sequentially suggests maximally informative experiments for improving this state of knowledge. In addition to contributing to the systems science knowledge base, the project will advance the state of the art in the fields of hierarchical Bayesian modeling, advanced inference methods, Bayesian model selection, and sequential experimental design. The research activities will enhance multidisciplinary collaboration between systems design and social science researchers, and facilitate the integration of theories and research methods from these disciplines. It will prepare graduate students with unique strengths at the interface of these fields. The results of the project will be disseminated through scientific publications, an open platform for deploying and executing experiments on human decision making, and computational tools for Bayesian analysis that will be distributed as open source software.