Genetic optimization with Dakota 2/3: A sandbox for experimentation

Part 1: The opti­miza­tion loop
Part 2: A sand­box for exper­i­men­ta­tion
Part 3: Practice

In order to play with the con­fig­u­ra­tion set­tings of Dakota’s moga opti­miz­er, you may use a “sand­box” demon­stra­tion opti­miza­tion case in which the results are quick­ly obtained and eas­i­ly visualized.

Attached is a demon­stra­tion opti­miza­tion case. A two-objec­tive, three-para­me­ter opti­miza­tion with exter­nal eval­u­a­tion (in Python) is con­fig­ured and exten­sive­ly com­ment­ed. Bash and Python scripts post-process the out­put of the sim­u­la­tion and dis­play what is hap­pen­ing (the per­for­mance of indi­vid­u­als in each gen­er­a­tion, and the size of the cor­re­spond­ing populations). 

Play­ing with the main set­tings of the muta­tion and crossover oper­a­tors, and watch­ing the results on the per­for­mance and pop­u­la­tion size graphs, is a great way to learn how they work.

Part 1: The opti­miza­tion loop
Part 2: A sand­box for exper­i­men­ta­tion
Part 3: Practice