Filter the controlled channel a little more (MAP) filter at 2 or 3.
Normal position is important to getting the last Nth of stable control.
Your WG system needs to be within its normal range of control for what you are targeting, otherwise controlling it is next to impossible, for instance if you used to use a standalone boost control with the same solenoid and your 250 kpaA target resulted in a flat 66% BCS DC the ECU controls will work fine, but if you were at 90% and had minimal control don't expect the ECU PID controls to fix anything.
Start with P of .1 leave I and D at 0.
Increase P .05 at a time until it overshoots target, add some D (.005 increments) until it is overshooting then oscillating to a stable setting, then add some I, say .01 at a try, stop at the minimal amount needed to eliminate overshoot. In some cases you may not need ANY Integral.
Hyperactive (wildly oscillating) BCS DC% and/or WG Position (if you have a sensor on it) are signs that one of the PID settings are way off. The top log graph with the wacko Aux 1 trace looks like one or more of the PID settings are an order of magnitude too high. The log posted by Martin shows a setup that is close to ideal for response.
If you are using MAP/EMAP for fuelling might suggest using Effcy on the X axis instead of RPM of the Normal Position table, as turbine relative backpressure (comp PR vs turbine ER) increases it will take more BCS DC% in nearly a direct ratio, you can pretty much guess what its going to need, low Effcy = more DC%.
Also keep in mind, especially for smaller turbos, once you overspeed the compressor any semblance of good boost control is out the window, don't fight a losing battle. For instance you have an 18 PSIg wastegate and above 6500 RPM it will only put out 14 PSIg manifold pressure at full BCS DC%. Well, its DONE regardless of what you do, you're at the end of the compressor map, don't fight it.