Bad handling of target values in scientific format

[codename5281]

Hello,
It would seem that during optimization values that are displayed in scientific format (i.e. -1.081e+0 instead of -1.081) are not properly handled during the optimisation phase. For example, i have the following iter table :

|   iter    |  target   | fruit_... | visibl... |
-------------------------------------------------
|  1        | -1.306e+0 |  41.7     |  1.108    |
|  2        | -80.06    |  0.01144  |  0.5233   |
|  3        | -85.11    |  14.68    |  0.2293   |
|  4        | -306.4    |  18.63    |  0.5838   |
|  5        | -997.5    |  39.68    |  0.8543   |
|  6        | -1.266e+0 |  41.92    |  1.059    |
|  7        | -711.4    |  20.45    |  1.329    |
|  8        | -65.8     |  2.739    |  1.039    |
|  9        | -1.081e+0 |  41.73    |  0.8822   |
|  10       | -138.0    |  14.04    |  0.3773   |
|  11       | -3.134e+0 |  80.07    |  1.456    |
|  12       | -940.9    |  31.34    |  1.069    |
|  13       | -3.259e+0 |  87.64    |  1.352    |
|  14       | -58.24    |  8.504    |  0.1547   |
|  15       | -582.6    |  16.98    |  1.329    |

in which best identified iter (in purple) will be iter 2, iter 8, and finally iter 14. However, do we agree that for a optimizer.maximize, the best target would be iter 9 ?
Thanks in advance,
Best,
J.

[bwheelz36]

Is this the same as #298?

[codename5281]

Yes, looks like it, didn't spot the other thread.
Closing the issue so the conversation stays in #298.