Distributed Biogeography-Based Optimization
Biogeography-based optimization (BBO) is an evolutionary optimization algorithm. Distributed BBO, referred to here as DBBO, is the extension of BBO to the situation when candidate solutions to an optimization problem cannot communicate with a central processor, and cannot communicate among themselves on a consistent basis. This situation arises in many practical applications, such as peer-to-peer networking, and robots’ whose communication is constrained by time or space.
The paper below proposes a DBBO algorithm, applies it to benchmark problems, and applies it to simulated and experimental robot controller optimization. The paper also derives a Markov model for DBBO. The MATLAB software that was used to derive the Markov model results can be downloaded in a zip file.
1. C. Scheidegger, A. Shah, and D. Simon, “Distributed Learning with Biogeography-Based Optimization: Markov Modeling and Robot Control,” submitted for publication - pdf, 258 KB
Carre Scheidegger ran the benchmark and robot control simulations, and wrote most of the paper. Arpit Shah designed and implemented the robot experiments. Carre and Arpit also wrote a preliminary conference version of this paper.
Last Revised: September 13, 2011