U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach," CRC Press, Boca Raton, FL, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....in Line (6) in Fig. 1. For the Max min heuristic, the completion time of a task is multiplied by .In the Sufferage heuristic, the sufferage value computed in Line (8) in Fig. 2 is multiplied by . 4. SIMULATION PROCEDURE The mappings are simulated using a discrete event simulator (e.g. [5, 14, 22]) The task arrivals are modeled by a Poisson random process. The simulator contains an ETC (expected time to compute) matrix that contains the expected execution times of a task on all machines, for all the tasks that can arrive for service. The ETC matrix entries used in the simulation studies ....

U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach," CRC Press, Boca Raton, FL, 1993.

....the minimax algorithm [13] Each Node is the result of actions taken by either of the participants that result in a significantly new state. States: A state is the minimal collection of information with which the system s future state can be uniquely predicted in the absence of chance events. [14] There 1 Normally in simulation literature, an event is that which causes a change in state. APSS is only concerned with significant changes in state, so Branches correspond more directly with transitions rather than with events, in the traditional simulation sense of event. are three kinds of ....

U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach, Boca Raton, FL: CRC Press, 1993.

....for each. A better approach, for this situation, is to segregate the simulation functionality from the functionality of each requirement. This prevents the duplication of effort at the simulator layer. In these systems the simulation layer has been implemented as a discrete event simulation [4]. A simulation executive (or SimExec) provides the user interface to set up and run the simulation. An event queue (SimEventQueue) maintains a list of simulation events (SimEvents) sorted by their execution time. Each SimEvent contains its execution time and a pointer to the event executor that ....

Pooch, U. W., and J. A. Wall. 1993. Discrete Event Simulation: A Practical Approach, CRC Press, Boca Raton, Florida.

....to each event definition. An instance is characterized by a timestamp and a list of parameters which depend on the event definition. Timestamps either already exist (for time events) or are generated for each event definition using modeling heuristics (like the discrete event technique [60]) All generated event instances are collected into a calendar which is a chronological list of future event occurrences. Instead of waiting for events to be signalled in the environment, the simulator picks one element after the other from the calendar and processes it. Concerning database state ....

U. W. Pooch, J. A. Wall. Discrete Event Simulation: A Practical Approach. CRC Press Boca Raton, Florida 93.

....in Line (6) in Figure 1. For the Max min heuristic, the completion time of a task is multiplied by z. In the Sufferage heuristic, the sufferage value computed in Line (8) in Figure 2 is multiplied by z. 4. Simulation Procedure The mappings are simulated using a discrete event simulator (e.g. [5, 14, 22]) The task arrivals are modeled by a Poisson random process. The simulator contains an ETC (expected time to compute) matrix that contains the expected execution times of a task on all machines, for all the tasks that can arrive for service. The ETC matrix entries used in the simulation studies ....

U. W. Pooch and J. A. Wall, "Discrete Event Simulation: A Practical Approach," CRC Press, Boca Raton, FL, 1993. 3

....to the APSS) The PE may launch a Planner at Node E to recommend more branches A B D C F EH G I Figure 4: Plan Description States: A state is the minimal collection of information with which the system s future state can be uniquely predicted in the absence of chance events. [14] There are three kinds of states maintained in this system: the Actual State, the Planned State, and the Anticipated State. The Actual State comes from the World View. A Planned State is generated when a Planner initially creates a Branch in the plan, and is held in a newly created Node in the ....

U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach. Boca Raton, FL: CRC Press, 1993.

....in state, so Branches correspond more directly with transitions composed of several actions rather than with the traditional notion of an event. A state is the minimal collection of information with which the system s future state can be uniquely predicted in the absence of chance events. [6] There are three kinds of states maintained in this system: the Actual State, the Planned State, and the Anticipated State. The Actual State comes from the World View. A Planned State is generated when a Planner initially creates a Branch in the plan, and is held in a newly created Node in the ....

U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach, Boca Raton, FL: CRC Press, 1993.

.... #) lower confidence interval for the proportion is p L = x x (n x 1)F 1 # 2 (r 3 ,r 4 ) 15) where F 1 # 2 (r 3 ,r 4 )isthe(1 # 2) quantile of the F distribution with r 3 =2# (n x 1) degrees of freedom for the numerator and r 4 =2# x degrees of freedom for the denominator [Hal52] [PW93]. 3 Numerical Results Implementing the interval estimators of sequential coverage analysis has been discussed in the previous section. Our other implementation rules for sequential coverage analysis on a single processor and multiple processors under MRIP (Multiple Replications In Parallel) ....

U. W. Pooch and J. A. Wall. Discrete Event Simulation: A Practical Approach. CRC Press, 1993.

.... In addition to the scheduling strategy, we use the Feasible Deadline (FD) and the Not Tardy (NT) criteria to decide whether a transaction should be processed [AGM92] The network delay was calculated as a constant object request transfer time plus a random exponentially generated value [PW93] The results that we present are: i) the percentage of transactions completed within their deadline in the two systems, ii) the cache hit percentages achieved in the CS RTDBS, and (iii) the percentage of object requests that could not be granted immediately by the server, in the CS RTDBS, due ....

U. Pooch and J. Wall. Discrete Event Simulation -- A Practical Approach. CRC Press Inc., Boca Raton, Florida, 1993.

....algorithms. In addition to the scheduling strategy, we use the Feasible Deadline (FD) and the Not Tardy (NT) criteria to decide whether a transaction should be processed [1] The network delay was calculated as a constant object request transfer time plus a random exponentially generated value [17]. The results that we present are: i) the percentage of transactions completed within their deadline in the two systems, ii) the cache hit percentages achieved in the CS RTDBS, and (iii) the percentage of object requests that could not be granted immediately by the server, in the CS RTDBS, due ....

U. Pooch and J. Wall. Discrete Event Simulation -- A Practical Approach. CRC Press Inc., Boca Raton, Florida, 1993.

....state, so Branches correspond more directly with transitions composed of several actions rather than with the traditional notion of an event. States A state is the minimal collection of information with which the system s future state can be uniquely predicted in the absence of chance events. [27] There are three kinds of states maintained in this system: the Actual State, the Planned State, and the Anticipated State. The Actual State comes from the World View. A Planned State is generated when a Planner initially creates a Branch in the plan, and is held in a newly created Node in the ....

U. W. Pooch and J. A. Wall, Discrete Event Simulation: A Practical Approach, Boca Raton, FL: CRC Press, 1993.

No context found.

U. W. Pooch and J. A. Wall, Discrete Event Simulation---A Practical Approach, CRC Press, 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC