Website of Randall J. Patton


Below are links to studies I have done on a range of topics, both during my working career and recently in retirement.  Full discussion for each topic is contained in the linked PDF files.

Long-range Torpedo Detection

This study showed the feasibility of detecting individual wake-homing torpedoes in a salvo at long distances by taking advantage of the torpedo's high speed and the frequency domain processing inherent in synthetic aperture sonar (SAS) processing. 

Survival Modeling

This study was inspired by a friend's observation that, the longer a person was a member of a group, the more likely they were to stay in the group. The ensuing analysis revealed a general technique for generating important probability distributions associated with survival modeling.


Science and Society

This paper describes how an innovative ocean acoustic experiment, aimed at providing data to analyze global warming effects, was stymied by ineffective communication between project scientists and sincere but ill-informed environmental activists.


Breaking the Sound Barrier

Felix Baumgartner broke the sound barrier in 2012 by jumping out of a balloon at 40 km altitude. Or did he? This study was inspired by the observation that, since both sound velocity and the terminal velocity of an object vary with altitude, it is unclear whether the object's speed actually exceeds the speed of sound. The analysis compares these altitude dependent functions using a realistic non-linear frictional force and a simple but realistic model of atmospheric profiles.

Ocean Swell Modeling

The movement of a submerged object caused by ocean swell waves is a well known phenomenon, especially to those with experience SCUBA diving. While models of ocean swell power spectra are readily available, the need arose to utilize these frequency domain representations to generate realizations in the time domain. This study presents a method of generating a time series of ocean swell based on the modulation of a sine wave using a Markov-process-based random time series. Oceanography meets AM radio.


Sales Predictions

A buyer needs to accurately predict the percent of stock left at 8 weeks based on sales from the previous 6 weeks. Simple extrapolation of a least-squares line doesn't work in this case because the shrinking amount of inventory reduces the amount available for sale. An exponential model, reflecting this connection between sales and inventory, does the trick. An example of an All Experts question.

Odd-shaped lot size

A landowner needs to know the size of her odd-shaped lot but the local surveyors can't deliver. All Experts to the rescue.

Please reload