Peter makes lots of sense. He has implemented a zero voltage crossing detector to find the "start" of the wave you are controlling, and then digitally proportions the firing of the triac with a burst of pulses, which are just couple in with a little pulse transformer.

Peter doesn't help much with your current problem ! His PSU design is fine, but if it is any distance from your micro, you are going to pick up huge amounts of radiated noise from your switched load. You need to put in a snubber across the load to kill the RFI. make sure that you have a very good earth to your board.

Steve