A Dynamic Programming Algorithm for Finding an Optimal Sequence of Informative Measurements

. 2023 Jan 30;25(2):251. doi: 10.3390/e25020251

Algorithm 5 Modified Stochastic Rollout for GPs

7:
for $i = k + 1$ to $N - 1$ do
8:
$σ_{i}^{2} (x_{0 : i}^{'}) = κ - p^{T} C^{- 1} p$
9:
${\hat{μ}}_{i}^{'} (x_{i}^{'}) \leftarrow G e n e r a t e_b a s e_p o l i c y (x_{i}^{'})$
10:
$w_{i} \sim p_{W_{i}}$
11:
$x_{i + 1}^{'} \leftarrow v_{i} (x_{i}^{'}, {\hat{μ}}_{i}^{'} (x_{i}^{'}), w_{i})$
12:
end for
13:
Store: $H_{k} (M_{k} | x_{0 : k}^{'}) + {\tilde{J}}_{k + 1} (x_{k + 1}^{'})$