Posted on: November 9, 2016 | J. Van Gurley and Lawrence D. Stone - Siam News

On July 23, 2013, the lobster boat Anna Mary was heading out from Montauk, New York, to the fishing grounds off Long Island. Around 9:00 p.m., Anthony Sosinski went below to get some sleep while his partner, John Aldridge, stayed on deck to prepare for the next day’s fishing. When Sosinski awoke at 6:00 a.m., the boat was still motoring out on autopilot but Aldridge was gone – no visible sign of when or where he fell overboard [9]. Given the many uncertainties in the circumstances under which Aldridge fell overboard, the motion of the boat, and the influence of tides, currents, and winds, how does one even begin to mount an effective search? The U.S. Coast Guard faces this type of scenario daily. The problem also occasionally arises for major aviation accidents, such as the disappearance of Air France Flight 447 over the South Atlantic in 2009 and Malaysian Air Flight 370 in the Indian Ocean in 2014. Fortunately, Bayesian search theory provides a principled approach for problems of this nature.

Developed by the U.S. Navy in response to the German submarine threat in the Atlantic Ocean during World War II, Bayesian search theory is a systematic mathematical method for planning searches for lost objects. It has been used to plan successful searches for lost submarines (USS Scorpion [5]), aircraft (Air France Flight 447 [7]), and treasure ships (SS Central America [6]). Bayesian search theory is the analytic core of the U.S Coast Guard’s national Search and Rescue Optimal Planning System (SAROPS), credited with helping save scores of lives, including that of John Aldridge.

More ›