"Discovery of Worst Case Go Sort Inputs"
summary
Implements a genetic algorithm in Go that treats each input array as a genotype = array and measures fitness by how poorly sort performs (swap count / runtime), evolves populations with crossover and mutation, and avoids contention by using random per goroutine RNGs and preallocated result slots to eliminate mutex bottlenecks. Builds and ships via Docker on AWS (Batch / ECS), stores results in DynamoDB, and successfully discovers pathological inputs that force Go’s sorter into its worst-case fallback behaviors.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:15What would you do
- 1:47Genetic Algorithms
- 5:07Sorting and Go
- 7:30Building a Genetic Algorithm
- 14:32AWS Components
- 15:51VPC Components
- 18:34AWS Documentation
- 20:33AWS Permissions
- 22:05AWS CLI
