We discussed on a future direction on his study. His SAD calculation should be interfaced to the BDSIM (based on GEANT4) in order to simulate background particles generated at collimators and its impact at detectors. Since probability of halo generation would be very small, particles must be tracked with weights in the simulation for enough statistical estimations. Technical details of how to realize this method in the SAD were also discussed.

In his simulation, mover correction has following 4 steps of (1) ground motion, (2) beam position measurement, (3) Q-mover correction and (4) accelerating structure (cavity) mover correction, and this process from (1) to (2) is repeated assuming there is no ground motion between (2) and (4). The correction is so called "simple correction". At the LINAC end, emittance was calculated at (1), (2) and (4). The ground motions were the GM-A,B,C of ILC-TRC report and KEK models which were presented at the ISG10 by Tauchi. At this simulation, no "pulse-to-pulse correction" was assumed.

First, he looked at emittance growth only with the ATL motion of A=1 x 10^{-17}m/s. The emittance was large after (3), i.e. Q-mover correction, since orbit changes by the Q-mover correction cause emittance dilution due to wakefield. This large emittance growth can be cured by the cavity mover correction.

Secondly, emittance growths with GM-A,B,C were calculated as a function of time(T_{cor}) between the corrections, i.e. a cycle time of the simple correction. Large emittance growth was obtained after ground motions of the GM-B and C at T_{cor} > 10s and >0.1s, respectively. This behavior shall require a fast (orbit) feedback in the LINAC. Since the GM-C has relatively rapid emittance growth even after the correction, T_{cor} > 10s will not be acceptable. He found that the very rapid growth ( T_{cor}>0.06s ) with the GM-C was due to elastic wave at 2.5Hz and a 50Hz wave had small effect.
He also showed an emittance growth with the GM of KEK-4am model which is the most quiet time. The result was between the GM-C and B. The KEK-10am ( the noisiest time) would be very similar to the GM-C. In these cases, even the GM-B, the "pulse-to-pulse" correction must be taken account.

Finally, he simulated a correction performance after beam shut downs of AT=1 x 10^{-10}m which corresponds to 3 months at A= 1 x 10^{-17} m/s. After the shut down, the emittance is worsen more than 5th order of magnitudes. However, the small emittance will be recovered only in 10 corrections.

He will investigate the KEK-model (especially the 10am model). The "pulse-to-pulse correction" shall be included in the SLEPT simulation. Results will be also presented in terms of luminosity as well as emittance growth.