Searching for an optimal solution is like hunting for pirate treasure on an island. You and your friends are the search agents, and every part of the book gives you a different way to search. Pick a hunting strategy, press Run, and watch it work on the same island.
Based on Alaa Khamis, Optimization Algorithms: AI techniques for design, planning, and control problems, Manning, 2024.
Every finished attempt is logged here, so you can compare strategies on the same island. "Spots checked" counts objective evaluations: how many places a hunter had to inspect.
| # | Strategy | Part | Outcome | Spots checked | Best elevation |
|---|
| On the island | In optimization |
|---|---|
| The whole island except the active volcano | Feasible search space |
| The active volcano (and the sea) | Infeasible region: violates a constraint |
| The treasure, marked X | The optimal solution (global optimum) |
| Elevation, because pirates hid treasure on high spots | The objective function to maximize |
| Lower hills such as Parrot Hill | Local optima that can trap a search |
| You and your friends | Search agents |
| Strength of the metal detector signal | Reward signal for learning-based methods |
| Records of earlier treasure hunts | Historical data to train a predictive model |
Fire Mountain is the highest point on the island, but it is an active volcano, so it can never be the answer. This is why constraints matter: the best feasible solution is not always the highest point overall.