Add more general methods for finding stable points

[Mv77]

Computing stable points in the ConsIndShockModel.py agents relies on finding expectations of future market resources (or ratios that involve them) given today's states (and optimal behavior). This operation is model-specific to the extent that each model has its own rules for how resources evolve (e.g., constant Rfree in IndShockConsumer and variable Rfrees in KinkedRconsumer).

Once HARK makes the move towards having explicit representations of transition equations, these model-specific methods should be susceptible of being abstracted. There could be a unified method that searches over states applying the transition equation --whatever it is-- to find next period's states until they satisfy the conditions required by each stable point.

Such method should be able to find stable points for the Kinked R model. See the discussion in ConsIndShockModel.py/ConsKinkedRSolver.addStablePoints() at #899