Good Lattice Rules in Weighted Korobov Spaces with General Weights

Abstract

We study the problem of multivariate integration and the construction of good lattice rules in weighted Korobov spaces with general weights. These spaces are not necessarily tensor products of spaces of univariate functions. Sufficient conditions for tractability and strong tractability of multivariate integration in such weighted function spaces are found. These conditions are also necessary if the weights are such that the reproducing kernel of the weighted Korobov space is pointwise non-negative. The existence of a lattice rule which achieves the nearly optimal convergence order is proven. A component-by-component (CBC) algorithm that constructs good lattice rules is presented. The resulting lattice rules achieve tractability or strong tractability error bounds and achieve nearly optimal convergence order for suitably decaying weights. We also study special weights such as finite-order and order-dependent weights. For these special weights, the cost of the CBC algorithm is polynomial. Numerical computations show that the lattice rules constructed by the CBC algorithm give much smaller worst-case errors than the mean worst-case errors over all quasi-Monte Carlo rules or over all lattice rules, and generally smaller worst-case errors than the best Korobov lattice rules in dimensions up to hundreds. Numerical results are provided to illustrate the efficiency of CBC lattice rules and Korobov lattice rules (with suitably chosen weights), in particular for high-dimensional finance problems.

The construction of good extensible rank-1 lattices

“…The proof of the second part is similar to the proof of a result in [6] which makes use of Jensen's inequality, namely that for a sequence of positive numbers…”

“…In the following we introduce the particular reproducing kernel Hilbert spaces in which numerical integration is frequently considered [3,5,6,9,13,14,16,17 [5] is given by…”

“…It follows from (3), (6) and (8) that (11) e n,s,2π 2 γ (P n,s (z); K s,2π 2 γ ) = e per,n,s,2,γ (P n,s (z); K per,s,α,γ ), where 2π 2 γ denotes the sequence of weights (2π 2 γ j ) j≥1 . Thus the results shown in the following are valid for the root mean square error for numerical integration in the Sobolev space as well as for the worst-case error for numerical integration in the Korobov space.…”

“…We now make a small adjustment to this set which allows us to incorporate Jensen's inequality; see [6] where a similar argument was used. For a real c ≥ 1 define the set…”

On the mean square weighted L2discrepancy of randomized digital (t,m,s)-nets over Z2

“…Including the lower dimensional projections much reduces this effect. The weights γ u are then introduced to modify the importance of the discrepancy of the projections, with the intention to adjust the measure to the usage of the point set (see [6,29]). For example it has been observed that in many applications the higher dimensional projections are considerably less important than the lower dimensional ones.…”