*1) Feature distance calculation*

The feature distance between **x**<sup>∗</sup> *qr* and each of the historical memory vectors in matrix **X**, is computed using the Manhattan distance (L1 -Norm) as

$$d\_i(\mathbf{X}\_i, \mathbf{x}\_{q\_r}^\*) = \left\| \mathbf{X}\_i - \mathbf{x}\_{q\_r}^\* \right\|\_1 = \sum\_{j=1}^p \left| \mathbf{x}\_{i,j} - \mathbf{x}\_{r,j}^\* \right| \tag{1}$$

and produces the distance vector, **d**∈ *<sup>M</sup>*�<sup>1</sup> .
