Greedy mind map
Make locally optimal choices for global solutions.
Core idea: Take a local move only when an exchange or stay-ahead argument says an optimum can include it.
Mnemonic: Prove safe → commit → never revisit · Cost: Usually sort O(n log n)
Order: Sort into safety
Signal: One ordering exposes a choice that leaves maximum room for the future.
Move: Sort by finish, deadline, or another proof-bearing key, then accept compatible choices.
Examples: Non-overlapping Intervals, Activity Selection, Queue Reconstruction
Extend: Best reachable frontier
Signal: Only the furthest or cheapest reachable state matters so far.
Move: Scan while maintaining the strongest frontier; fail when the scan outruns it.
Examples: Jump Game, Gas Station, Partition Labels
Pair: Commit one side
Signal: Sorted extremes can be matched and one side cannot benefit from waiting.
Move: Pair when possible; always advance the endpoint whose fate is already determined.
Examples: Boats to Save People, Assign Cookies, Two City Scheduling
Schedule: Best available resource
Signal: Jobs arrive over time and a heap can choose among currently legal options.
Move: Sort releases, add available work, then pop the locally safest job or resource.
Examples: Task Scheduler, Single-Threaded CPU, Meeting Rooms