Reminder: This post contains 471 words
· 2 min read
· by Xianbin
When designing randomized algorithms, there are many ways. The following paradigms in [1] are useful.
- Fooling an adversary
- Random Sampling
- Abundance of witness
- Fingerprinting and hashing
- Random re-ordering
- Load Balancing
- Rapidly mixing Markov chains
- Isolation and symmetry breaking
- Probabilistic methods and existence proofs.
Paradigms 2, 6, 7, 9 can be seen a lot in papers. In this blog, I will cover all the above methods by examples.
Reference
[1]. Motwani, Rajeev, and Prabhakar Raghavan. Randomized algorithms. Cambridge university press, 1995.