My Linux rigs all auto restart in about a min if they drop a GPU. This is relatively straight forward in Linux, and both optiminer and Claymore provide watchdog capabilities and scripts that can be implemented via cron. I prefer to do this from a control PC, ssh, and the json port for the miner so that one PC (raspberry Pi) controls multiple rigs and can even kill power to a rig if needed (via Zwave and OpenHAB). Its a bit overkill but I travel allot and sometimes a hard power reset can save the day when I am away (beats trying to talk the Wife through reseting a miner).
I have just started testing the EWBF miner and am disappointed that there is no watchdog capabilities, but it does have a json port (that I don't have working yet). I started testing in Windows but the Nvidia drivers allow over-clocking directly in Linux unlike AMD, so I have switched to Linux. I am finding the Linux miner is much more stable.
I am working on a way to auto restart with the EWBF miner in Linux but since you mention a .bat file I assume you are working in Windows. I do not know an easy way to do what you want in Windows, others may be of more help. If you switch to Linux let me know.