What strategies from queuing theory can manage traffic flow better?

Automotive are an exemplar queue-making solution. In fact, never designed with queues in mind. Queues generate from the combination of: 1) Ability of units to move faster than the optimal capacity for the moment 2) Information travelling backward slower than the 1) speed. Plus a very important factor: 3) Drivers zigzagging between lanes, causing everyone in the the queue to press on forward. From that the answer is primarily dual: - dense enough (e.g. every 100m) real time display of the optimal local speed - strongly preventing lane switch in a queue.

Beware that local optimization quickly become counter-productive at a wider area scale. No dynamic optimization should be considered bottom-up (local first) but always in a top-down manner, at traffic jam-free bounadries

Remember that automotive traffic is intrinsically generating disorder, because local decision (interest) is disconnected from the collective optimal. Any urban area wide control system would need to control effectively the mobile units, which could go as far as remote control w/o driver intervention. ...then this is no longer automotive, but fine grained train to the absurd (in cost, safety, efficiency etc.)

The CarLina system of city-car sized, automatic guided transport for people and goods, implements the best combination of jam-free urban area. Not by resorting to fancy technologies nor expensive, complex control, but to the opposite, by eliminating everything that would make the automotive necessary in an urban area: full area deterministic scheduling to start with, in a dedicated, elevated network with zero interaction to ground or other stuff. Once managed for optimal operation, the system proves vastly more efficient, inexpensive, available, flexible than automotive, inclufing delivery in an urban area. Everything is solved at the system architecture level = the very opposite of cars.

Queuing theory offers several strategies to optimize traffic flow, such as implementing traffic lights with optimal cycle lengths and timings, designing efficient lane configurations to minimize congestion, and employing advanced traffic signal control algorithms like SCATS (Sydney Coordinated Adaptive Traffic System). Additionally, dynamic traffic management systems utilize real-time data to adjust signal timings and lane assignments, reducing delays and improving overall traffic efficiency. Moreover, strategies like congestion pricing can incentivize drivers to adjust their travel times, alleviating peak-hour traffic congestion and balancing demand across different times of the day.