“…To maximize utility in expectation, we employed dynamic programming [Bellman, 1957], a technique that solves complex stochastic optimization problems by breaking them into simpler, solvable subproblems [Dasgupta et al, 2006]. Dynamic programming has been used to solve decision problems under weather uncertainty in fields including power dispatch [Hable et al, 2002], irrigation [Wilks and Wolfe, 1998], and air traffic management [Nilim et al, 2001]. Using dynamic programming, the implications of any flight decision can be broken into two parts: the implications for today and the implications for the rest of the experiment [Hanlon et al, 2013].…”