Innovation Path: Discovering an Ordered Set of Optimized Intermediate Solutions from an Existing to a Desired Solution
Ahmer Khan, Kalyanmoy Deb · Proceedings of the Genetic and Evolutionary Computation Conference · 2024
In practice, there is often a need to modify an existing implemented solution in order to achieve a better performance or accommodate new demands or technologies. However, a new optimal solution for the updated problem may be quite different from the existing solution, thereby causing an apathy for its implementation involving large cost, changes, and efforts. For such scenarios, we propose a concept of an "innovation path" (IP) containing a sequence of intermediate solutions from the existing to the new target solution with gradual and controlled change from one to the next. To discover intermediate solutions of the IP simultaneously, we propose a bi-objective formulation of the original problem, so that Pareto-optimal solutions of the resulting bi-objective problem become the IP solutions. We demonstrate the working of the proposed approach on a number of single and two-objective test and engineering problems. Results are encouraging and suggest further research and application to make the proposed innovation path approach more efficient and practical.