Local search using global views for large scale combinatorial optimization

摘要

Problem to be solved: to provide local search using global view for large scale combinatorial optimization.How to solve large combinatorial optimization problemsBy at least one processorEnter the objective function and initial solution as mapping from multiple N nodes andMultiple nodesRandomly clustering the K clusters of N / K nodes, respectively.For each cluster of K clustersAssigning binary variables to each possible permutation of the label set in the cluster andIf u > 2, u = K2 variableThen, it is determined whether u = 1 variable exists in the case of k = 2 andThe purpose is to represent the function for the UN / K variable andUsing the quadratic constraint free binary optimization (Qubo) solver to obtain an update function for obtaining an update solution for the UN / K variableUpdate solutionDetermine whether the convergence criteria are met or not.For the determination that the convergence criterion is satisfiedOutput an update solution for the purpose function and.Diagram