I am trying to justify the choice of state for my QL formulation. Intially the state only consisted of: {a single quality level to represent the quality of the best solution, the computed arcs, the current location}. This seemed to be obfuscating some states, and I modified my formulation to be {quality level for path along top arc, quality level for path along bottom arc, minimum arc, the computed arcs, the current location}.
I feel that this new formuation does not blend as many underlying states together at the risk of a larger dimentional state space. I have a diagram of the possible choices of combinations of state parameters and plan to discuss them.