Convert BDD strategy to dictionary strategy
As we discussed in the meeting, so we can use etcsim.py for the HSCC paper.
Because the BDD strategy is nondeterministic, we need to determinize it to get a proper dictionary mapping states to sampling decisions. My suggestion is to sample as late as possible, i.e., if a state admits both sample and wait actions, choose sample. Of course some cases will have equivalently late options, such as "loop 1 trigger and loop 2 wait" versus "loop 1 wait and loop 2 trigger"... in that case we can arbitrate for now (or choose whatever heuristic rule, like trigger the one with highest clock valuation).