This paper integrates an interactive framework into Automatic Construction technique of Loading Algorithms based on Cartesian Genetic Programing (ACLA-CGP). While ACLA-CGP has a capacity to derive a loading algorithm optimizing a given objective function, its validity still remains unclear under human evaluations. Our interactive ACLA-CGP, i.e., i-ACLA-CGP, is designed to reduce human evaluations as possible and thus evaluates’ burdens. Experimental results show that i-ACLA-CGP successfully derives loading algorithms with 50 human evaluations while satisfying their preferences.