Goheannee
(2014)
Performances Of Metaheuristic Algorithms In Optimizing Tool Capacity Allocations.
Masters thesis, Perpustakaan Hamzah Sendut.
Abstract
Semiconductor manufacturing industry in general has moved into high mix productions
resulting from the drastic pace of product innovation. Capacity planning In
semiconductor manufacturing facility, such as allocating right mix of products to
maximize the capacity output, needs to consider multiple mutually influenced constraints
in resource, product demand, as well as product and process characteristics. To achieve
the best allocation, optimization methods, such as metaheuristic algorithms are
commonly used. This research compares the performances of various metaheuristic
algorithms to optimize tool capacity allocation in two case studies. In this research, the
algorithms studied includes Genetic Algorithm, Particle Swarm Optimization Algorithm,
Differential Evolution Algorithm, Harmony Search Algorithm, Teaching-LearningBased
Optimization Algorithm and Black Hole Algorithm. These algorithms are inspired
by different nature of phenomenon. The former three are common in literature for tool
capacity allocation problems. The latter three are the next generation of metaheuristic
algorithms and albeit popular elsewhere, have no known attempt in tool capacity
allocation problems. The case studies were obtained from two real industries and five
demand scenarios were derived. The demand scenarios were with different demand
intensities and levels. For each case study, a capacity model was constructed in
Microsoft Excel spreadsheet, as an input to the above mentioned metaheuristic
algorithms which programmed in Matlab coding. The performances considered are tool
utilization and aggregate capacity outputs.
Actions (login required)
 |
View Item |
