Does the temporal difference learning of optimal utility values (U) require knowledge of the transition probability tables

No

Explanation:

Temporal difference or some times written as TD learning process may be defined as an approach to learning that describes how to predict a given  quantity which depends on the future values for a given signal.

TD or temporal difference learning does not require the knowledge of transition probability tables. It only requires the knowledge of state and action plan. It also does not require the knowledge of reward function.

answer: a. the kind of population living in your neighbourhood.

the census explorer data tool by the us census bureau allows the user to zoom in on any neighbourhood within the united states and find the kind of people residing in that by race and age.  

the tool needs to needs to be used twice – once in order to find what type of people reside in a neighbourhood, and then to see the age groups of the people in the same neighbourhood.