liveBook · Manning

Pheromone evaporation

As explained in Algorithms Every Programmer Should Know

An ant colony optimization operation that reduces pheromone weights across graph edges to weaken previously favored paths.

In ant colony optimization, pheromone evaporation reduces the pheromone weight on every graph edge by a chosen percentage. This negative feedback weakens old paths and helps prevent the swarm from getting stuck on a mediocre route. section 8.1 section 8.4

Halving the weights

With a 50% decay rate, every edge starting at 1.0 falls to 0.5 after evaporation. In a 60-mile tour example, traversed edges then receive a separate deposit, reaching 2.17; ignored edges remain at 0.5. section 8.1

Figure 8.3 The graph state after one complete ant tour. Traversed edges (solid lines) receive a pheromone deposit, compounding their weight to 2.17. Ignored edges (dotted lines) suffer from mandatory evaporation, reducing their weight to 0.5. Over multiple iterations, the dotted lines will evaporate entirely to zero.

The graph shows traversed edges gaining pheromone while ignored edges lose it through evaporation; repeated updates can reduce ignored paths further. section 8.1

How Algorithms Every Programmer Should Know covers it

How deep it goes
Defines graph-wide percentage decay and traces a 50% update; explains its role in preventing premature convergence.
Examples
Uses a 60-mile tour to illustrate halved edge weights and describes pheromone decay shifting network traffic to a backup route.
What it assumes
Recognizes pheromone weights and completed ant tours; explains evaporation within the optimization feedback loop.
What you can do after
Understand how graph-wide decay weakens ignored paths and reduces the risk of the swarm settling too early on a poor route.
How it connects
Connects evaporation with pheromone deposit, ant colony optimization, and probabilistic path selection.

Also explained in

Compare how 2 books explain pheromone evaporation →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Algorithms Every Programmer Should Know. The section links point to the parts of the book it draws on.