How to response to those 3 following questions

Rational Agent ( Artificiall intellegence)

1. There can be more than one agent program that implements a given agent function.

(If you agree, provide an example; otherwise argue/demonstrate why it is not possible.)

2. There are agent functions that cannot be implemented by any agent program.

(Hint: what if the agent function required deciding an undecidable problem1?)

3. Consider what would occur if we kept an agent program fixed, but sped up the machine upon which it was run by a factor of two - does that change agent behavior?

(Hint: consider properties of the environment with which the agent is tasked.)