Abstract
This paper considers the parallel implementation of a genetic algorithm using a network of transputers. The context of the work is the design of a flight control system using a novel approach to Robust Eigenstructure Assignment. The robustness properties of the Linear Quadratic Regulator are exploited and Eigenstructure Assignment is performed to achieve desired aircraft handling qualities by projecting the desired eigenstructure into an output feedback constrained LQR subspace using a cost function minimisation approach. Both the motivation for parallelisation and the actual transputer implementations are considered. Indications of scalability and complexity are discussed in the light of specific implementation issues.
Original language | English |
---|---|
Pages (from-to) | 11-19 |
Number of pages | 9 |
Journal | Control engineering practice |
Volume | 3 |
Issue number | 1 |
Publication status | Published - Jan 1995 |
Keywords
- GENETIC ALGORITHM
- TRANSPUTER
- PARALLEL PROCESSING
- EIGENSTRUCTURE ASSIGNMENT