Routing algorithms which do not base their routing decisions on measurements…

Concept

Routing algorithms are classified by whether they react to live network conditions. A non-adaptive (static) routing algorithm fixes its routing decisions in advance and never changes them in response to the current traffic load or topology. An adaptive (dynamic) routing algorithm, by contrast, continuously measures or estimates the present network state and updates its routes accordingly.

Applying it here

The stem describes an algorithm that does NOT base its decisions on measurements or estimates of the current traffic and topology. By definition, an algorithm that ignores the live network state and uses precomputed, unchanging paths is a non-adaptive (static) routing algorithm.

• Precomputed paths: routes are calculated offline and loaded into the routers when the network boots.

• Static behaviour: because it does not sense current conditions, it cannot react to congestion or link failures in real time.

• Simplicity: easy to implement and well suited to small or stable networks where traffic and topology rarely change.

Why the other categories do not fit

• Adaptive (dynamic) routing is the exact opposite: it DOES use live traffic and topology measurements to update routes, so it cannot describe an algorithm that ignores them.

• Greedy is a general algorithm-design strategy (make the locally optimal choice at each step); it is not the standard category used to classify routing algorithms by their use of network state.

• Back propagation is a training method for neural networks, unrelated to how routers choose paths.