Interdisciplinary Problems and Numerical Analysis : 10 Things We Wish We

Non-engineering faculty often find the engineering quantitative mindset and ability to conduct numeric analysis helpful in their research, yielding valuable results not otherwise discernible by either specialty alone. New engineering faculty can find such work helpful to launch their careers by exposing them to a wealth of productive research topics relatively untouched by single-discipline researchers, as well as providing opportunities to get to know many faculty and be exposed to a variety of research methods, writing styles, and grant sources. Despite the demonstrable benefits of such collaborations there also are pitfalls, especially for new engineering faculty who have little experience coordinating complex interdisciplinary projects. This paper describes observations on interdisciplinary collaborations based upon and referenced to several dozen interdisciplinary papers the authors have published with faculty from clinical medicine, bioengineering, finance, educational psychology, colonial history, business, sports medicine, and seismology. The paper includes five reasons to seek opportunities to apply numerical analysis to interdisciplinary problems, three common pitfalls of work in such interdisciplinary projects, and ten best practices for conducting numerical analysis of interdisciplinary problems. I. Reasons to seek interdisciplinary numerical analysis opportunities Interdisciplinary research often reveals low-hanging fruit As a graduate student, one of the authors was the lone electrical engineer in a biomedical center that had a predominantly molecular chemistry emphasis. His specialty was analog hardware design, but naturally anything that had to do with signal analysis, data analysis, or more generally mathematics fell on his desk as well. As he worked with the various biology postdoctoral scientists he found that several relatively elementary engineering analysis techniques could be used to solve problems they encountered in novel ways, such as the use of finitedifference techniques to predict fluid flow 1 , or strain fields in arterial tissue 2 , or using relatively simple geometrical relationships to experimentally measure arterial surface strain 3 . The experimental papers that resulted were relatively fast to write since they were in largelyuninvestigated fields, in contrast to the depth required to publish in journals of comparable impact factor in his relatively well-researched area of sensor design 4 . This is a well-known relationship; the authors of popular Teaching Engineering text observe: “It is easiest to get results and write publications when you work on new ideas instead of following the well-beaten research track 5 .” The ability to speed-publish such papers is especially important for new engineering professors given the reality that the tenure process places enormous pressure on new hires to publish quickly and in quantity 6,7,8,9 .