»ã±¨±êÌâ (Title)£ºPontryagin¡¯s maximum principle, optimal control problem solver RIOTS_95, machine learning and fractional calculus (ÅÓÌØÀïÑǽ𼫴óÖµµÀÀí£¬×îÓŽÚÔìÎÊÌâÇó½âÆ÷RIOTS_95, »úе½ø½¨ºÍ·ÖÊý½×΢»ý·Ö)

»ã±¨ÈË (Speaker)£º³ÂÑôȪ ½ÌÊÚ£¨University of California, Merced, USA£©

»ã±¨¹¦·ò (Time)£º2023Äê12ÔÂ1ÈÕ(ÖÜÎå) 11:00

»ã±¨µØÖ· (Place)£ºÌÚѶ»áÒé(363-278-420)

Ô¼ÇëÈË(Inviter)£ºÀƷ¡¢²ÌÃô

Ö÷°ì²¿ÃÅ£ºÀíѧԺÊýѧϵ

»ã±¨ÌáÒª£ºThis seminar is of tutorial nature to inspire research efforts in connecting optimal control (OC) and machine learning (ML) and fractional calculus (FC). First, a tutorial on static and dynamic optimization problems is presented with a uniform Lagrangian multiplier framework. After a simple exposure to calculus of variation (COV) and its transversality conditions, we derive the basic solution based on Hamiltonian and the Pontryagin¡¯s maximum principle (PMP). An analytical simplest example is presented to show the need of a general-purpose toolbox to solve OCPs (optimal control problems). Then RITOS_95, a general OCP solver in the form of a Matlab toolbox is introduced with some solved example OCPs. Live OCP solving runs in Matlab will be shown.