Memory Design for Constrained Dynamic Optimization Problems

A proposal for a memory design is given that is suitable for solving constrained dynamic optimization problems by an evolutionary m. Based on ideas from abstract memory, two schemes, blending and censoring, are introduced and tested. Using a new benchmark we show in numerical experiments that such a memory can improve solving certain types of constrained dynamic problems.

• Type: Conference
• Year: 2010
• Where: EVOW
• Authors: Hendrik Richter
• Keywords: Abstract Memory, Artificial Intelligence, Constrained Dynamic Optimization, Constrained Dynamic Problems, EVOW 2010