Abstract:We present streaming algorithms for the graph $k$-matching problem in both the insert-only and dynamic models. Our algorithms, with space complexity matching the best upper bounds, have optimal or near-optimal update time, significantly improving on previous results. More specifically, for the insert-only streaming model, we present a one-pass algorithm with optimal space complexity $O(k^2)$ and optimal update time $O(1)$, that with high probability computes a maximum weighted $k$-matching of a given weighted graph. The update time of our algorithm significantly improves the previous upper bound of $O(\log k)$, which was derived only for $k$-matching on unweighted graphs. For the dynamic streaming model, we present a one-pass algorithm that with high probability computes a maximum weighted $k$-matching in $O(Wk^2 \cdot \mbox{polylog}(n)$ space and with $O(\mbox{polylog}(n))$ update time, where $W$ is the number of distinct edge weights. Again the update time of our algorithm improves the previous upper bound of $O(k^2 \cdot \mbox{polylog}(n))$. This algorithm, when applied to unweighted graphs, gives a streaming algorithm on the dynamic model whose space and update time complexities are both near-optimal. Our results also imply a streaming approximation algorithm for maximum weighted $k$-matching whose space complexity matches the best known upper bound with a significantly improved update time.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to efficiently find the Maximum Weighted k - Matching in the graph stream model. Specifically, the paper focuses on how to design stream algorithms to achieve the optimal or near - optimal update time in the insert - only model and the dynamic model, while keeping the space complexity matching the existing best upper bound.
### Main contributions of the paper
1. **Contributions in the insert - only model**:
- A one - pass randomized stream algorithm is proposed, which can construct the maximum weighted k - matching in a weighted graph.
- The space complexity of the algorithm is \(O(k^2)\), and the update time is \(O(1)\), both of which are optimal.
- Compared with the previous best result (the deterministic kernelized stream algorithm of Fafianie and Kratsch, with space complexity \(O(k^2)\) and update time \(O(\log k)\)), this algorithm not only achieves the same optimal space complexity but also reaches the optimal update time \(O(1)\).
- In particular, this algorithm can handle the k - matching problem of weighted graphs, which is more challenging than handling the k - matching problem of unweighted graphs.
2. **Contributions in the dynamic model**:
- A one - pass randomized stream algorithm is proposed, which can construct the maximum weighted k - matching in a weighted graph.
- The space complexity of the algorithm is \(O(k^2W\cdot \text{polylog}(n))\), and the update time is \(O(\text{polylog}(n))\), where \(W\) is the number of different edge weights in the graph.
- This result directly implies a one - pass randomized stream algorithm for the k - matching problem of unweighted graphs, with near - optimal space complexity \(O(k^2\cdot \text{polylog}(n))\) and near - optimal update time \(O(\text{polylog}(n))\).
- Compared with the previous best result (the algorithm of Chitnis et al., assuming that the maximum matching size of the graph does not exceed \(k\), with space complexity \(O(k^2\cdot \text{polylog}(n))\) and update time \(O(\text{polylog}(n))\)), this algorithm does not rely on this strong assumption, and without this assumption, the update time is significantly improved to \(O(\text{polylog}(n))\).
3. **Contributions of the approximation algorithm**:
- A one - pass stream approximation algorithm is proposed, which can construct a \((1 - \epsilon)\) - approximate solution of the maximum weighted k - matching in a weighted graph stream with high probability for any \(\epsilon> 0\).
- The space complexity of the algorithm is \(O(k^2\log R_{\text{wt}}\cdot \text{polylog}(n)/\epsilon)\), and the update time is \(O(\text{polylog}(n))\), where \(R_{\text{wt}}\) is the ratio of the maximum edge weight to the minimum edge weight in the graph.
- This result improves the update time complexity of the previous approximation algorithm (the same space complexity, but the update time is \(O(k^2\cdot \text{polylog}(n))\)).
### Technical details
- **Compact Subgraph**: The key observation in the paper is that there is a compact subgraph of size \(O(k^2)\) that contains the maximum weighted k - matching in the original graph. By using universal hashing techniques, this compact subgraph can be efficiently identified.
- **Many - to - Many Mapping**: To achieve a faster update time, the paper