By Eugen Lamers

For the simulation of stochastic platforms, these recommendations are crucial that are in a position to accelerate the simulation, i.e. the identical of accelerating the accuracy that may be reached inside of a given run-time, specifically with a spotlight on infrequent events.

Eugen Lamers explains the main of simulation speed-up regularly, demonstrates the procedure RESTART for the simulation of infrequent occasions, and addresses its potency on allotted platforms. as a result personality of RESTART, the potency is proscribed by means of the excessive conversation overhead. the writer introduces the non permanent Dynamic Simulation notion, constructed for the making plans of cellular radio networks, which allows quick kingdom area assurance in addition to the dignity of dynamic points. He analyses the speed-up power on a theoretical foundation with analytically solvable Markov types for various utilisations. With lifelike types for UMTS community making plans he indicates that the assessment of dynamic functionality measures advantages from the temporary Dynamic Simulation procedure.

Show description

Read Online or Download Contributions to Simulation Speed-Up: Rare Event Simulation and Short-Term Dynamic Simulation for Mobile Network Planning PDF

Best research books

Looking toward the future of technology-enhanced education: ubiquitous learning and the digital native

Taking a look towards the way forward for Technology-Enhanced schooling: Ubiquitous studying and the electronic local bridges the space among know-how and schooling by means of proposing leading edge examine at the way forward for schooling. a necessary reference on e-learning, this scholarly ebook examines present examine in expertise more suitable studying, presents new didactic versions for schooling, and discusses the latest applied sciences and their influence on schooling.

Optimization Theory, Decision Making, and Operations Research Applications: Proceedings of the 1st International Symposium and 10th Balkan Conference on Operational Research

Those complaints encompass 30 chosen study papers in keeping with effects offered on the tenth Balkan convention & 1st overseas Symposium on Operational examine (BALCOR 2011) held in Thessaloniki, Greece, September 22-24, 2011. BALCOR is a longtime biennial convention attended via various college, researchers and scholars from the Balkan nations but additionally from different ecu and Mediterranean international locations besides.

Additional resources for Contributions to Simulation Speed-Up: Rare Event Simulation and Short-Term Dynamic Simulation for Mobile Network Planning

Sample text

25 for g = 4. The first step in the interval [0,I0 ] occupies in such cases the greatest portion of the total simulation time. This is why the efficiency for g = 4 is significantly higher than for g = 2 in the case of the lower load η = e−4 . For the higher load, the distance G(I0 ) for the first possible threshold I0 is closer to the optimum than for the lower load case, which leads to a smaller reduction of the sample size for g = 4 compared to g = 2. This reduction is too small to compensate the higher computation effort for the higher granularity, which results in lower total efficiency.

The most relevant attributes of these possible choices are the scalability and the communication overhead. 3. Here, they are discussed with respect to their applicability to RESTART. 3. For this approach, model components have to be identified to decide about the way they will be assigned to processes. These components need to be as autonomous as possible to minimise the necessary communication between the processes. Components can be the network nodes, preferably including the associated queue.

1, the basic equation is PB = PB|I · PI . 1 RESTART fundamentals 23 ..... G0 = 1 ..... G1 ..... Gj H0 = G I0 j −1 j 0 1 Step 0 n0 Step 1 n1 I0 ..... 2 shows on the x-axis the value space of the random variable from x = 0 to x = B. Larger values are not included since the value at x = B represents P{X ≥ B}, and for cases in which values of x > B are possible, all values x ≥ B are considered as the rare event set, and there is no concern for further fragmentation of the area x ≥ B. The example value space displayed in the figure is discrete leading to the stepped distribution function.

Download PDF sample

Contributions to Simulation Speed-Up: Rare Event Simulation by Eugen Lamers
Rated 4.17 of 5 – based on 35 votes