| Summary: | 碩士 === 國立中央大學 === 土木工程學系 === 84 === The traditional transportation problem is solved with linear
costs inorder to reduce the complexity of the problem size. In
many instances,with the increasing of unit goods, the marginal
transportation costs aredecreased with the delivery levels and,
therefore, the objective functionsare concave.This research thus
aims to develop heuristic methods to solvethe concave cost
transportation problems. The main solution method is based on
the combinatorial optimizationheuristic techniques, such as
Simulated Annealing,Threshold Accepting andTabu Search. The
relation between transportation problem and
combinatorialoptimation were discussed. From the point of view
of the accepting rate,the heuristic methods suitable for this
problem were developed and improvedin different ways. A network
generator was designed to test the parametersin different
heuristic methods. Many instances were generated to teston an
HP workstation by using the C computer language. The
performancebetween different algorithms were discussed. This
research developed agood initial solution method to improve the
solution quality. We foundthat Threshold Accepting-based
algorithms performed relatively well in thetests.
|
|---|
