Implement CaseBase as a graph structure where similar cases are clustered together. For retrieval 1) Use Random Multistart 2) List of K size for chosen cases 3) Threshold values to prevent returning a dissimilar case 4) Greedy algorithm