A digression around VA's SRC #012b

11282022, 02:11 PM
Post: #21




RE: A digression around VA's SRC #012b
(11272022 03:45 PM)Albert Chan Wrote: FNROOT is interpolating the secant line for root, from points (x1,y1), (x2,y2) My mistake. From HP Journal, July 1984 Quote:The search for a root proceeds by bentsecant, quadratic interpolation, and minimabounding methods, depending on the detected shape of the function. FNROOT always return something. Confirmation of root, FVALUE ≈ 0, is required. Quote:FNBOOT will return a value when it finds a root, a local minimum of absolute value, or a pair of neighboring machinerepresentable numbers across which the function changes sign, when the search leads out of the range of representable numbers, or when quadratic interpolation has been used eight times in succession. 

« Next Oldest  Next Newest »

User(s) browsing this thread: