Abstract
Many real-world design problems involve posing and solving an optimization
problem. The process of optimization aims at achieving the best possible
result(s) under given circumstances. Very often, the design problem
under consideration is sufficiently complex mathematically and cannot
be solved to optimality by most classical (gradient-based) optimization
algorithms. To alleviate the difficulties faced by the gradient-based
optimization algorithms, several non-traditional optimization algorithms
have been proposed and successfully used to solve such design problems.
The non-traditional optimization algorithms can handle complexities
such as simultaneous optimization of multiple objectives, multi-modal
function profiles, non-convex and discontinuous feasible search spaces,
and mixed variables. Evolutionary algorithms, simulated annealing,
particle-swam optimization, ant-colony based algorithms are some
of the non-traditional optimization algorithms derived from Nature.
One of the major drawbacks of most non-traditional methods is the
high computational cost (large number of function evaluations) incurred
during the process of optimization. Reducing the computational cost
incurred and improving the robustness of the optimization algorithms
while still maintaining the desired performance characteristics is
the motivation behind the proposed research.
Users
Please
log in to take part in the discussion (add own reviews or comments).