Search Problem More Interesting

[Sungonogi]

Currently the search problem transition model is sometimes not very interesting. It can happen that

• an action is never applicable
• some states are not reachable
• some states are dead ends, even non goal states

• also there is currently always only one goal state. One could try to also allow two goal states, but depending on the amount of states it might make it impossible to have solutions with some minimum Length (f.e. if there are only 5 states and we have two goal states and an initial state, then there are no solutions that take more steps than inital -> nongoal1 -> nongoal2 -> goal). If you add this then as parameter numGoalStatse chosen in the generateOne function to make sure it is actually possible

So think about possible improvements, but be careful that all combinations are possible to generate and take <1s (use the testAllCombinationsValid function)