Carl Boettiger

Thurs­day, August 15 8:00 Cof­fee & Break­fast at NIM­BioS 8:45 Train­ing Prob­lem Groups TASK: Pre­pare progress report/action strat­egy for full group 9:30 Regroup: Reports from each Train­ing Prob­lem Group 10:30 Plan­ning for Meet­ing 2 12:00 Lunch at NIM­BioS & Departure

Tools Shar­ing: Drop­box (Team account? 350 gigs/12 users $1670/yr) Com­mon lan­guage for train­ing devel­op­ment: Mat­lab Remote col­lab­o­ra­tion: GoToMeet­ing? Evo? Code col­lab­o­ra­tion GitHub? Train­ing prob­lem I 1a) State vars > Con­trol vars: Can give solu­tions with chat­ter­ing (peri­odic, pos­si­bly jump­ing between bound­aries.) Short-term pri­or­i­ties: 1) shared numer­ics.  Jim looks at col­lo­ca­tion prob­lem: how do they per­form on peri­odic con­trols …

Wednes­day, August 14 8:00 Cof­fee & Break­fast at NIM­BioS 8:45 Intro­duc­tion to Day 3 (Donahue/Armsworth) 9:00 Train­ing Prob­lem Groups Marie-Jose reflects on the sam­pling prob­lem: Opti­miza­tion between effort and data. Spa­tial auto­cor­re­la­tion chal­lenges peri­odic sam­pling chal­lenges. Fortin et al 1989 Carl asked: Nyquist fre­quency? Spa­tial sam­pling the cor­rect scale by max­i­mum vari­ance in wavelet scal­ing. Frank reflects …

Tues­day, August 13 8:00 Cof­fee & Break­fast at NIM­BioS 8:45 Out­line for Day 2 (Donahue/Armsworth) Adap­tive Con­trol, Jake Intro­duced by he two-armed ban­dit prob­lem, Rothchild i = 1,2. \( T_i \) is the num­ber of tri­als, \( N_i \) are num­bers of suc­cesses. Define \[ \rho_i = \frac{1}{1+T_i}, \mu_i = \frac{N_i}{1+T_i} \] If i is played …

Mon­day, August 12 8:00 Cof­fee & Break­fast at NIM­BioS 8:45 Wel­come to NIM­BioS (Lou Gross, Direc­tor of NIM­BioS) 9:00 Intro to the PDG Con­trol Work­ing Group (Paul Armsworth) 9:30 Intro­duc­tions by each Work­ing Group Par­tic­i­pant 5-minute self-introduction to the group, includ­ing your back­ground, exper­tise, and cur­rent work 10:30 Break 11:00 Intro­duc­tion to Train­ing Prob­lem 1 …