Abstract
It's intractable to implement parallel genetic algorithm (PGA). By analyzing the framework of MapReduce and characteristics of Island PGA, using Mappers as Islands, Reducer administers the whole algorithm evolution, proposed Island MapReduce-based Parallel Genetic Algorithm (IMRPGA). The advantages of IMRPGA are: 1) Comparing with traditional PGAs, the implement of IMRPGA is simple, it doesn't need to be concerned about that the synchronization, dealing with communication errors caused by parallelism; 2) It is easy for IMRPGA to expand to very large scale. This paper showed the detail design, analyzed the theoretical speed-up and implemented IMRPGA on Hadoop platform. Verified the feasibility of IMRPGA and its speed-up, the results indicate that IMRPGA suits to solve the time-consuming problems and the complex object functions.
Original language | English |
---|---|
Pages (from-to) | 486-493 |
Number of pages | 8 |
Journal | International Review on Computers and Software |
Volume | 6 |
Issue number | 4 |
Publication status | Published - Jul 2011 |
Externally published | Yes |
Keywords
- Cloud computing
- Island mode
- Mapreduce
- Parallel genetic algorithm