Abstract:We study the weighted star discrepancy of the Halton sequence. In particular, we show that the Halton sequence achieves strong polynomial tractability for the weighted star discrepancy for product weights $(\gamma_j)_{j \ge 1}$ under the mildest condition on the weight sequence known so far for explicitly constructive sequences. The condition requires $\sup_{d \ge 1} \max_{\emptyset \not= \mathfrak{u} \subseteq [d]} \prod_{j \in \mathfrak{u}} (j \gamma_j) < \infty$. The same result holds for Niederreiter sequences and for other types of digital sequences. Our results are true also for the weighted unanchored discrepancy.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the tractability of the weighted star discrepancy of Halton sequences. Specifically, the author studies under what conditions the Halton sequences can achieve strong polynomial tractability, especially in the case of using product weights. ### Problem Background 1. **Weighted Star Discrepancy**: The weighted star discrepancy is a measure for evaluating the uniformity of point - set distribution, especially suitable for high - dimensional spaces. It takes into account the importance of different dimensions and adjusts the influence of each dimension through weights. 2. **Halton Sequences**: Halton sequences are a class of low - discrepancy sequences, often used in quasi - Monte Carlo methods to improve the accuracy and efficiency of numerical integration. 3. **Tractability**: In computational complexity theory, tractability refers to whether the complexity of an algorithm can be kept within an acceptable range as the problem size (such as the number of points $N$ and the dimension $d$) increases. Strong polynomial tractability means that the complexity grows polynomially with $N$ and $d$ and is independent of the dimension $d$. ### Main Contributions of the Paper - **Weakest Conditions**: The paper shows the weakest conditions for Halton sequences to achieve strong polynomial tractability under product weights. Specifically, if the following condition is satisfied: \[ \sup_{d \geq 1} \max_{\emptyset \neq u \subseteq [d]} \prod_{j \in u} (j\gamma_j) < \infty, \] then the weighted star discrepancy of Halton sequences can achieve strong polynomial tractability, and the $\varepsilon$-exponent is at most 2 at this time. - **Improved Results**: The paper also proves that if a more stringent condition is satisfied: \[ \sum_{j \geq 1} j\gamma_j < \infty, \] then the weighted star discrepancy can achieve the optimal $\varepsilon$-exponent of 1, that is, the strong polynomial tractability is optimal. ### Conclusion Through strict mathematical derivation and analysis, this paper provides the currently weakest conditions for the strong polynomial tractability of Halton sequences under weighted star discrepancy to be achieved. This is of great significance for high - dimensional numerical integration and other related applications, especially providing more relaxed requirements in terms of weight selection. ### Formula Summary - Definition of Weighted Star Discrepancy: \[ D^*_{N, \gamma}(P_d) := \sup_{\alpha \in [0,1]^d} \max_{\emptyset \neq u \subseteq [d]} \gamma_u \left| \Delta_{P_d}((\alpha_u, 1)) \right| \] - Upper Bound of the Weighted Star Discrepancy of Halton Sequences: \[ D^*_{N, \gamma}(H_{b_1, \ldots, b_d}) \leq \frac{1}{N} \max_{\emptyset \neq u \subseteq [d]} \prod_{j \in u} (6j\gamma_j \log N) \] - Conditions for Strong Polynomial Tractability: \[ \sum_{j \geq 1} j\gamma_j < \infty \quad \text{or} \quad \sup_{d \geq 1} \max_{\emptyset \neq u \subseteq [d]} \prod_{j \in u} (j\gamma_j) < \infty \]

Tractability properties of the weighted star discrepancy of the Halton sequence

Scrambled Halton Subsequences and Inverse Star-Discrepancy

Probabilistic weighted star discrepancy bound based on the general equal measure partition

Conditions for tractability of the weighted $L_p$-discrepancy and integration in non-homogeneous tensor product spaces

Characterization of the convergence of weighted averages of sequences and functions

A rigorous computer aided estimation for Gelfond exponent of weighted Thue-Morse sequences

On the regularization of sequences and associated weight functions

Lebesgue constants for the Walsh system and the discrepancy of the van der Corput sequence

On Weighted Zero-Sum Sequences

Varieties Of Regularities In Weighted Sequences

Weighted Ehrhart Theory: Extending Stanley's nonnegativity theorem

Smooth discrepancy and Littlewood's conjecture

String matching with swaps in a weighted sequence

Polynomial p-adic Low-Discrepancy Sequences

Metrical lower bounds on the discrepancy of digital Kronecker-sequences

Zeros distribution and interlacing property for certain polynomial sequences

On strong growth conditions for weighted spaces of entire functions

Smooth weighted zero-sum constants

Strong tractability of multivariate integration using quasi-Monte Carlo algorithms

A constructive approach to strong tractability using Quasi-Monte Carlo algorithms

Some properties of discrete classes of weights satisfying reverse Hölder inequality and reverse inequality for nonincreasing sequences