J. S. Stadler and S. Roy, "Adaptive importance sampling," IEEE Journal on Selected Areas in Communications, vol. 11, no. 3, pp. 309--316, April 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....estimator gives no information at all, from the Information Theory point of view, which means (again) that we can get the answer only if we already knew it beforehand. People, nevertheless, have been using this ideal estimator as a guideline to find practical estimators [1] 11] 12] 13] [14]. It is up to the reader to believe in its usefulness (we do not) and try it out. 3.2. Increasing the probability of the event to be analyzed Some authors have suggested (implicitly or explicitly) that a safe way of finding an efficient twisted density function is by increasing the probability ....

....the original distribution. 3.4. Adaptive Importance Sampling Adaptive Importance Sampling is a technique in which, instead of basing the selection of the twisted density function on analytical or theoretical results, it is done by sequentially improving the estimator s performance [4] 9] [14]. In other words, the estimator is evaluated using a certain set of statistical measures for its performance (which could be the sample variance of the estimate, some other scattering measure, and or even the estimate itself) and then its twisted density function is modified according to a ....

Stadler, J. S., Roy, S., "Adaptive Importance Sampling", IEEE Journal on Selected Areas in Communications, Vol. 11, No. 3 (April), pp 309-316, 1993.

....defined by a single hyperplane, or halfspace. As discussed in [12] the decision boundary for the equalization problem is a nonconvex union of partial half planes, and hence the linear shift method performs poorly in such applications. An IS simulation method using a recursive adaptation technique [20] to update the simulation density has also been proposed. However, this method cannot guarantee the convergence of simulation density parameters for our SBSD structure. An alternative technique based on large deviation theory (LDT) 21] is developed in [22] Under certain geometric conditions on , ....

J. S. Stadler and S. Roy, "Adaptive importance sampling," IEEE J. Select. Areas Commun., vol. 11, pp. 309--316, Apr. 1993.

.... error floor induced by the random phase drift; from this information we design an IS procedure based on large deviations theory (LDT) Our references, for this purpose are [2] 3] and [4] When departing from the asymptotic error floor an adaptive importance sampling (AIS) approach, adapted from [5], is adopted. The paper is organized as follows: Section 2 presents the adopted communications model and some IS aspects relevant to the simulation design. In Section 3 we derive the error set for density biasing using LDT, and present the main aspects behind the AIS technique applied. ....

....# =## # . Optimization of IS consists now of biasing in the product space #####. For minimization of # ## we use a stochastic search, because we have no information about the error set. The error set is now E ## # ######. We use a parametric AIS technique adapted from that proposed in [5]. We estimate the conditional mean E #### ;V # ;I# # ### ;V # ;I# # E ## # # . The error set is conditioned on the transmitted symbol. Optimization must yield a multiple bias solution that will constitute the bias for density p # ### ;V # ;I#. We avoid repeating here the details in ....

[Article contains additional citation context not shown here]

J. S. Stadler and S. Roy, "Adaptive importance sampling," IEEE Journal on Selected Areas in Communications, vol. 11, no. 3, pp. 309--316, April 1993.

....the theory of Importance Sampling (IS) IS involves twisting the PDF of the model output to increase the probability of the event under investigation, this means that the simulation time can be reduced. Chen et al. 27] present a paper on IS applied to communication systems while Stadler and Roy[28] discuss Adaptive IS (AIS) AIS does not need to know the PDF of the traffic because the algorithim changes as traffic is produced. The ability of IS to improve estimation of the CLR has important implications for ATM control. Call Admission Control (CAC) can be implemented using IS because an ....

J. S. Stadler and S. Roy. Adaptive importance sampling. IEEE Journal on Selected Areas in Communications, 11(3):309--315, 1993.

No context found.

J. S. Stadler and S. Roy, "Adaptive importance sampling," IEEE Journal on Selected Areas in Communications, vol. 11, no. 3, pp. 309--316, April 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