AI SearchAn adaptive hybrid algorithm for multi-mode resource-constrained project scheduling problems
Call Number: AIT Thesis no.ISE-09-04 Material type:
SeriesSeries: Asian Institute of Technology. Thesis ; no. ISE-09-04Publication details: Pathum Thani, Thailand : Asian Institute of Technology, 2009Description: 92 p. : illSubject(s): Online resources: Dissertation note: Thesis (M.Eng.) - Asian Institute of Technology, 2009 Summary: This thesis focuses on the multi execution modes resource-constrained project scheduling problem with the objective of minimizing the project makespan. There might be more than one operating mode to perform an activity; each mode requires different amount of resources and related time duration. Besides, there are two types of resources: renewable and non-renewable. Adaptive particle swarm optimization and genetic algorithm are integrated in an algorithm called APSO-GA to solve the MRCPSP. The major motivation of APSO-GA is to use PSO to find the best priority of activities while GA is used to search for the combination mode of the activities. A potential solution of MRCPSP is represented as a pair of particle and chromosome and an active schedule can be achieved by transforming this representation by serial schedule method. In adaptive PSO algorithm, the two parameters evolved in velocity updating mechanism 12,cc can be self-adaptive depending on three factors, including their old values, the degree of acceleration and the swarm response. At the beginning of the searching procedure, the cognitive learning term plays a more important role than the social learning term in effect to particle{u2018}s velocity. The effect of cognitive learning term is decreased while that of social learning term is increased through the PSO iteration. Genetic algorithm is put inside the PSO algorithm, i.e. in each PSO iteration, a GA loop is used to search for a better mode combination of activities. The GA loop is performed until the exploring process cannot find a better combination of mode for current activities priority arrangement. Moreover, three difference types of GA crossover variants are applied to enhance the quality of searching procedure. The performance of the APSO-GA algorithm is investigated on standard instance sets. Experiment results show that the proposed adaptive PSO algorithm has proved the advantages over the non-adaptive one and the best GA crossover types is uniform crossover. The makespan results and computational time results of the proposed algorithm are also very competitive in comparing with others heuristics previously published.
| Cover image | Item type | Current library | Home library | Collection | Shelving location | Call number | Materials specified | Vol info | URL | Copy number | Status | Notes | Date due | Barcode | Item holds | Item hold queue priority | Course reserves | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
22-AIT Thesis (Replacement)
|
Asian Institute of Technology Library AIT Publications | AIT Thesis no.ISE-09-04 (Browse shelf(Opens below)) | 3 | Available | 30050120700934 | |||||||||||||
40-Archives
|
Asian Institute of Technology Library Archives | AIT Thesis no.ISE-09-04 (Browse shelf(Opens below)) | 1 | Available | 30050160030689 |
Submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Industrial and Manufacturing Engineering, School of Engineering Technology
This thesis focuses on the multi execution modes resource-constrained project scheduling problem with the objective of minimizing the project makespan. There might be more than one operating mode to perform an activity; each mode requires different amount of resources and related time duration. Besides, there are two types of resources: renewable and non-renewable. Adaptive particle swarm optimization and genetic algorithm are integrated in an algorithm called APSO-GA to solve the MRCPSP. The major motivation of APSO-GA is to use PSO to find the best priority of activities while GA is used to search for the combination mode of the activities. A potential solution of MRCPSP is represented as a pair of particle and chromosome and an active schedule can be achieved by transforming this representation by serial schedule method. In adaptive PSO algorithm, the two parameters evolved in velocity updating mechanism 12,cc can be self-adaptive depending on three factors, including their old values, the degree of acceleration and the swarm response. At the beginning of the searching procedure, the cognitive learning term plays a more important role than the social learning term in effect to particle{u2018}s velocity. The effect of cognitive learning term is decreased while that of social learning term is increased through the PSO iteration. Genetic algorithm is put inside the PSO algorithm, i.e. in each PSO iteration, a GA loop is used to search for a better mode combination of activities. The GA loop is performed until the exploring process cannot find a better combination of mode for current activities priority arrangement. Moreover, three difference types of GA crossover variants are applied to enhance the quality of searching procedure. The performance of the APSO-GA algorithm is investigated on standard instance sets. Experiment results show that the proposed adaptive PSO algorithm has proved the advantages over the non-adaptive one and the best GA crossover types is uniform crossover. The makespan results and computational time results of the proposed algorithm are also very competitive in comparing with others heuristics previously published.
Thesis (M.Eng.) - Asian Institute of Technology, 2009
There are no comments on this title.