Certify greedy strategies with exchange argument proofs and optimal substructure.
npx skills add https://github.com/Arcadi4/nerdy --skill greedy-algorithms
Please help me install this Agent Skill. Skill: greedy-algorithms Source: https://github.com/Arcadi4/nerdy/tree/main/clrs Command: npx skills add https://github.com/Arcadi4/nerdy --skill greedy-algorithms
Greedy algorithms provide a disciplined framework for obtaining fast, often optimal, solutions by making locally optimal choices and proving safety.
Validate a proposed greedy rule by identifying the greedy choice, showing the residual subproblem remains of the same form, and proving safety with an exchange argument.
High-intent search queries and answers about installing and using this skill.
To prove a greedy algorithm is correct, identify the greedy choice, show the residual subproblem retains the same structure, and demonstrate safety by swapping an optimal solution's choice with the greedy choice without degrading the overall result.
Use a greedy algorithm instead of dynamic programming when locally optimal choices guarantee global optimality, verifiable via exchange arguments and optimal substructure. Dynamic programming is preferred when overlapping subproblems exist without a provably safe greedy choice.
The greedy approach for activity selection picks the earliest finishing interval to maximize count, while fractional knapsack selects items by highest value-to-weight ratio. Both require exchange arguments to prove the locally optimal choice yields a globally optimal solution.
Write a proof skeleton for Huffman coding by establishing optimal substructure, then apply an exchange argument demonstrating that merging the two lowest-frequency nodes is a safe choice that exists within some optimal prefix code tree.
A greedy algorithm produces a suboptimal solution for offline caching when the local eviction rule lacks safety. You must verify the greedy decision via an exchange argument to ensure the local choice matches the optimal global strategy.
Arcadi4Define states and reconstruct solutions for optimization problems using dynamic programming.
JamkrisMeasure performance baselines and detect regressions across code changes.
KentoShimizuFormalize states, transitions, and base cases for dynamic programming solutions.
rnavarychDesign algorithms with formal Big-O complexity analysis and correctness proofs.
