| Budget Amount *help |
¥4,110,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥510,000)
Fiscal Year 2007: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
Fiscal Year 2006: ¥1,900,000 (Direct Cost: ¥1,900,000)
|
|---|
| Research Abstract |
Quantum computer is a computation model using quantum mechanical principles such as superposition state, interference effect, and entanglement state. Recently, stochastic combinatorial search algorithms combined with evolutionary algorithm have been recently proposed by incorporating quantum mechanical principles or quantum bits. Narayanan, et. al. have proposed Interference Crossover (IX) for Classical Genetic Algorithm (CGA) in Traveling Salesman Problem (TSP), and have shown that IX can reduce search cost to 2/3 in CGA with a problem involving 9 cities. We have also shown that the combination of IX and Immune Algorithm (IA) shows better search performance than classical IA in TSP problems involving more than 50 cities.
Han, et. al. have proposed Quantum-inspired Evolutionary Algorithm (QEA) in which each gene is represented by a quantum bit. QEA can do single-point search and automatically shift from global search to local search like Simulated Annealing (SA). QEA can also perform mu
… More
lti-point search like CGA in order to solve large-scale optimization problems. In QEA, there are more than one subpopulations (groups) like Island GA (IGA), and inter- and intra-group migration procedures are performed. Evolution in each group enables coarse-grained parallelization and prevents premature convergence, and the migration procedures can control search diversification and intensification. However, the adjustment of a number of parameters is required for the number of group and migration intervals for each problem. In fact, Han, et. al. had to do vast experiments in order to get guidelines for the parameter adjustment in KP.
In this research, we propose a simpler algorithm which is referred to as Quantum-inspired Evolutionary Algorithm with Pair-Swap strategy (QEAPS). QEAPS involves just one population and a simple genetic operation which exchanges each best solution information between two individuals chosen randomly. Therefore, QEAPS involves less parameters necessary to be adjusted than QEA. We evaluate the search performance of QEAPS on 0-1 Knapsack Problem (KP), and show that QEAPS can find similar or even highly qualified solutions more efficiently and stably than QEA. Less
|
