Scheduling of Multi-stage Batch Processes Using Dynamic

Scheduling of Multi-stage Batch Processes Using Dynamic Decomposition on a Grid ƒ Computer manufacturers focus on multi-core architectures; computational resources become cheap ƒ Algorithms that utilize new architectures or exploit parallelization necessary ƒ Batching/scheduling problems have hierarchy of decisions Selection of tasks ⇒ Unit-task assignments ⇒ Task sequencing ƒ Exploit hierarchy to generate subproblems spawned to workstations ƒ Hard subproblems are dynamically decomposed ƒ Successful implementation using grid computing P

Master computer: Problem decomposition

Task selection Workers: Solution of subproblems Solutions/bounds collected by master

P1

P2

P3

PM1

P4

Unit-task assignment Master collects unsolved problems; 2nd-level decomposition

P12

P22

PM2 2(2)

P14

P24

PM4 2(4)

Task sequencing Process repeated as necessary

2 P1,1

2 P1,2

P12,M 3

4 P2,1

4 P2,2

P24,M 3

Strong branching