Pro-actively Interactive Evolution for Computer Animation

This paper addresses the issue of allowing a user more control over the outcomes of interactive evolution for computer animation, a variation on genetic algorithms. By using a point-for-point representation for genotype, the control both in selection and variation process is possible while the genotype represented by procedural rules in previous works allows it only in the selection. We experimentally validate these points on generating different walk styles out of a prototype walk motion and also discuss about interactive evolution to be used as a general approach to setting parameters for computer graphics.