Global optimization methods in computational chemistry

Question

I'm looking for a current and comprehensive overview (like a review article) of global optimization methods and their application in computational chemistry. Mostly I'm interested in geometry optimization. I am also grateful for single recommendations including

• literature (articles, textbooks)
• programs, capable to perform global optimizations
• presentations

In this case it would be very nice, if you could assign the methods to certain areas like solid state, molecules on surfaces, isolated molecules, ensembles, etc. .

Answer 1

You say you're mostly interested in geometry optimization. From a software point of view, here are a couple to get you started.

1. USPEX, which works with crystals, isolated molecules and nanoparticles, using a genetic algorithm and linking with many DFT and empirical codes from MATLAB (e.g. VASP, CASTEP, Siesta, GULP...). The link also has good articles summarising the technique and comparing it to some alternatives method for particular test cases, including a comparison for both crystal systems and Lennard-Jones clusters.

2. Constrained Minima Hopping, implemented here in the Atomistic Simulation Environment, so again can link with your DFT/MD code of choice. This is aimed at adsorbates on a surface, and you can find a good description in this paper by A. Peterson.

As I mentioned, there are links to review papers from both of the above methods, which should help you on your way to a more comprehensive literature search (e.g. more review articles).

It's worth noting that there are also these options:

3. Ab-Initio Random Structure Searching, which I don't know much about.
4. The Birmingham Cluster Genetic Algorithm (BCGA), which I guess is similar to the USPEX approach above, only aimed specifically at clusters.