INPUT: Size of each view L, Pc: probability of crossover, Pm: probability of mutation, K: number of views to be selected and G: the maximum number of generations |
Core: Step 1: Initialization which generates randomly the population of the Top-K views as PTKV. Step 2: Divide a PTKV into two equilibrium sub-populations PTKV1 and PTKV2 with size equal to PTKV/2. Step 3: Compute the fitness for PTKV1 and PTKV2 using CMV and CNMV respectively as a two bi-objective optimization problem. Step 4: Select the Top-K views from PTKV1 and PTKV2 using the proportionate selection based on the fitness of the Top-K views, and Compute subpopulation of Top-K views SPTKV1 and SPTKV2 using the corresponding objective function. Step 5: Combine the Top-K views in SPTKV1and SPTKV2 and generate a mating population of Top-K views MPTKV. Step 6: Apply the modified crossover and the random mutation operators to generate the offspring population of the Top-K views. |
OUTPUT: Top-K views TKV |