Greedy Random: A Novel Algorithm for Vehicle Routing Optimization
Research Purpose & Rationale
- To determine what contributes to Greedy Random (GR)'s success. We do this so that we can build better heuristics for HuGSS by building upon the success of something that works.
Contents
- Abstract - The abstract for this research.
- Research Paper - The actual paper and related items.
- Titles - Thoughts on a title for the research...
- Research Plan - What this research attempts to accomplish.
- Timeline - A log of what's been done and what remains to do.
Experiments
Some files linked to in the experiments below do not exist -- they took up too much web hosting space. But the experimental summaries are all there.
- Steepest Greedy Random - What if GR is really just doing very quick steepest searchs after an initial random move? Final Results (8/8 benchmarks).
- Variable Priorities Long GR - Same as variable priorities... but with a 90 minute run. Final Results (8/8 benchmarks).
- Minimize-Routes GR - What if GR is run with minimize routes as its objective function? Final Results (8/8 benchmarks).
- Feasible/Infeasible GR - What if GR is constrained to making only feasible or infeasible moves? Final Results (8/8 benchmarks).
- Variable Priorities GR - What happens if the moved customer is left at high or low, rather than medium? Final Results (8/8 benchmarks).
- Multi-move GR - We want to find if there is a "drop-off" point where GR makes "too many" random moves to find improvements. Considers 1, 2, 5, 7, 10, 25 and 50 moves. Final Results (8/8 benchmarks).
- Ideas - Ideas for experiments, also containing information about experiments yet-to-be-run.
- Procedures - General procedures we use to run experiments.
Specifications
- Metrics - How can we tell a heuristic is better, and what is it we want to record?
- StatAnalyzer - A modifed form of Data Analyzer fine-tuned to ease statistical analysis and conform with our new metric system.
- Heuristics - Specifications for each of the heuristics used in this study.
Presentations
- Greedy Random - Powerpoint presentation for the National Junior Science & Humanities Symposium.
- Pre-GR - LaTeX presentation for the Research Science Insitute 2000.
Links
