Design by derivational analogy: Issues in the automated replay of design plans

Abstract

Derivational analogy solves a problem by replaying the plan used to solve a previous problem, modifying it where necessary. We analyze how four published systems use this approach to help design (or redesign) complex artifacts like programs and circuits. We compare how they represent, acquire, and retrieve design plans; how they determine which parts of the old and new designs correspond; how they decide which steps of a design plan are appropriate to replay and adapt them to the new problem; and how they reuse partial plans. We show how each system's approach to these seven issues affects the SCOPE of problems it can solve, its EVOLVABILITY to solve new problems, the QUALITY of its solutions, the EFFICIENCY of its computation, and its AUTONOMY from the user.