Parallel simulation for business process re-engineering
Behrouz Zarei
Abstract
Behrouz Zarei
Abstract
This paper describes the development of a parallel simulation model for analysis of a business process system. Today’s increasing complexity of such systems, together with the need for more accurate results means that the task of simulation of business processes frequently outgrows the computational power of economically viable single process systems. One solution is to partition the computational task into sub-tasks and implement them on a parallel or distributed processing computers. This paper describes the process of modelling and execution of a simulation model in the context of a database development process and investigates the performance of the parallel execution of the model. The simulation computes the staff utilisation in various departments and the throughput of the system. The result of the parallel execution demonstrates that conservative parallel simulation algorithm using deadlock avoidance can provide significant speedup over a sequential simulation. Also the results of some experiments indicate that lookahead plays an important role in the performance of the simulation and, therefore, selecting the right lookahead generation technique can make a substantial difference. Key-Words: Discrete-event simulation, Parallel discrete event simulation, Lookahead, Business process simulation, Database management system
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper describes the development of a parallel simulation model for analysis of a business process system. Today’s increasing complexity of such systems, together with the need for more accurate results means that the task of simulation of business processes frequently outgrows the computational power of economically viable single process systems. One solution is to partition the computational task into sub-tasks and implement them on a parallel or distributed processing computers. This paper describes the process of modelling and execution of a simulation model in the context of a database development process and investigates the performance of the parallel execution of the model. The simulation computes the staff utilisation in various departments and the throughput of the system. The result of the parallel execution demonstrates that conservative parallel simulation algorithm using deadlock avoidance can provide significant speedup over a sequential simulation. Also the results of some experiments indicate that lookahead plays an important role in the performance of the simulation and, therefore, selecting the right lookahead generation technique can make a substantial difference. Key-Words: Discrete-event simulation, Parallel discrete event simulation, Lookahead, Business process simulation, Database management system
Key concepts: Computer science, Speedup, Process (computing), Deadlock, Parallel computing, Discrete event simulation, Task (project management), Partition (number theory)