Meta-Heuristics in Multi-Core Environments

CPU manufactures have been adding more cores to CPU's instead of only focusing on increasing the speed. There are many interesting questions regarding implementation and performance of algorithms using the new CPU's. We look at three well known meta-heuristic algorithms with different ratios between independent calculations and shared memory usage and analyze the benefits of using multi-core CPU's compared to single core CPU's. By programming specifically for multi-core processors, the performance of meta-heuristic algorithms can be improved without much effort, but the improvement is heavily dependent on the system architecture and the ratio between independent calculations and shared memory usage.