What is the difference between exploitation and exploration in the context of optimization?

In the paper Moth-flame optimization algorithm: A novel nature-inspired heuristic paradigm (2015, published in Knowledge-Based Systems)

The test functions are divided to three groups: unimodal, multi-modal, and composite. The unimodal functions ($$F1 - F7$$) are suitable for benchmarking the exploitation of algorithms since they have one global optimum and no local optima. In contrary, multi-modal functions ($$F8 - F13$$) have a massive number of local optima and are helpful to examine exploration and local optima avoidance of algorithms

I imagine that exploration means it goes searching for something in unknown regions from a starting point. But exploitation would search more around the starting (or current point).

It is more or less that? What else differentiates both concepts?

• In the specific context of reinforcement learning (which can also be thought of as a collection of optimization techniques), exploitation means "to choose the best current action", while exploration means "to choose a random action". However, in your context, these terms seem to be used slightly differently. However, to provide a reasonable answer to your question, we should also know what a test function is in this context. Either you explain it, or we need to read the paper. – nbro Nov 8 '20 at 11:25
• Test functions are fitness functions, such as the sum of (Xi)², going from 1 to n,. This if f1 for instance. From F1 to F7, all are sum of Xi after doing something else with it. All except F6, that it is the max value of the |xi| from 1 to n. – user2752471 Nov 10 '20 at 22:58