The bottleneck
Binary heap decrease-key = O(log V). E decrease-keys → O(E log V). Fib heap collapses this to O(E).
Advertisement
Fibonacci heap essentials
Lazy structure: consolidate only on extract-min. Trees of arbitrary rank. Amortized bounds via potential function.
Advertisement
Complexity
Extract-min V times: O(V log V). Decrease-key E times: O(E). Total: O(E + V log V).
Practical reality
Constants are HUGE. Fibonacci heaps rarely beat binary heaps in practice for typical graph sizes. Binary heap wins to N ≈ 10⁷.