Algorithm 1 Policy and Model Optimization (PMO)

1:Input: ${\mathcal{D}}_B$ and ${\mathcal{D}}_E$                                                                   Stage 01 2:Output: ${\tau }_{\theta }$ 3:$f_{{\varphi }_0}\left(\tau \right|\mathcal{X})\leftarrow {\mathcal{D}}_B$                                                                       Stage 02 as per Equation (1) 4:Initialize ${\pi }_{\theta }$, ${\mathcal{D}}_{\pi }$, $L_E$ 5:while not done do: 6:      Episode = sample(${\mathcal{D}}_E$) 7:      for $i=0:L_E$ do: 8:             $(\mathrm{state}:x_{t_c},\mathrm{action}:f_{{\varphi }_i}\left(x_d\right))\leftarrow f_{{\varphi }_i}\left({\tau }_{{\varphi }_i}\right|x_d)$ 9:             Append: ${\mathcal{D}}_i\leftarrow \langle x_{t_c},{\tau }_{{\varphi }_i}\rangle$ 10:           ${\mathcal{D}}_B\leftarrow {\mathcal{D}}_B\cup {\mathcal{D}}_i$ 11:           ${\mathcal{D}}_{\pi }\leftarrow {\mathcal{D}}_{\pi }\cup {\mathcal{D}}_i$ 12:           Every $n_b$   and   ${\mathcal{D}}_{\pi }\ge 2L_E$ do: 13:                 Compute loss for DBM: MSE (${\mathcal{L}}_{\varphi }$) 14:                 DBM Optimization: $f_{{\varphi }_i}\leftarrow {\mathcal{L}}_{\varphi }$                             Stage 02 on repeat 15:                 Compute Behavioral loss: (${\mathcal{L}}_B$)                           From Equations (3) and (4) 16:                 Compute policy loss: (${\mathcal{L}}_{\theta }$)                                    From Equations (3) and (4) 17:                 Policy Training & Optimization: ${\pi }_{\theta }\leftarrow {\mathcal{L}}_{\theta }$        Stage 03