Split `SimulationAgentLogic` into different traits

[paulheinr]

Currently, the SimulationAgentLogic consists of more and more functions necessary for any part of the simulation. I think it makes sense to split the functions into separate traits. E.g.

fn curr_act(&self) -> &InternalActivity;
fn next_act(&self) -> &InternalActivity;
fn curr_leg(&self) -> &InternalLeg;
fn next_leg(&self) -> Option<&InternalLeg>;
fn advance_plan(&mut self);
fn state(&self) -> SimulationAgentState;

Are general logic functions that should stay in this trait, but

fn wakeup_time(&self, now: u32) -> u32;
fn curr_link_id(&self) -> Option<&Id<Link>>;
fn peek_next_link_id(&self) -> Option<&Id<Link>>;

are either leg or activity related and should not be part of the same trait.