Norihito Yasuda

Maximizing hosting capacity of distributed generation by network reconfiguration in distribution system

The maximization of distributed generation (DG) hosting capacity that takes into account network configuration is a complex, non-linear combinatorial optimization problem. The search space of the configurations becomes massively large in practical-size networks with several hundreds of switches. For this reason, no existing method can handle such large-scale networks. In this paper, we propose a novel exact solution method. Our method consists of two stages. In the first stage, the method divides the entire problem into a set of small subproblems. In the second stage, it converts all subproblems into a compressed data structure called a zero-suppressed binary decision diagram (ZDD), which expresses the combinatorial sets compactly. The proposed method avoids any combinatorial explosion by using the ZDD to enable operations of the weighted combinatorial item sets. We conducted experiments on a large-scale network with 235 switches. As a result, our method obtained the global optimal solution in 49 hours.

A Dynamic Programming Algorithm for Tree Trimming-based Text Summarization

Tree trimming is the problem of extracting an optimal subtree from an input tree, and sentence extraction and sentence compression methods can be formulated and solved as tree trimming problems. Previous approaches require integer linear programming (ILP) solvers to obtain exact solutions. The problem of this approach is that ILP solvers are black-boxes and have no theoretical guarantee as to their computation complexity. We propose a dynamic programming (DP) algorithm for tree trimming problems whose running time is O(NLlogN), where N is the number of tree nodes andL is the length limit. Our algorithm exploits the zero-suppressed binary decision diagram (ZDD), a data structure that represents a family of sets as a directed acyclic graph, to represent the set of subtrees in a compact form; the structure of ZDD permits the application of DP to obtain exact solutions, and our algorithm is applicable to different tree trimming problems. Moreover, experiments show that our algorithm is faster than state-of-the-art ILP solvers, and that it scales well to handle large summarization problems.

Distribution Network Verification for Secure Restoration by Enumerating All Critical Failures

If several feeders are interrupted in a severe accident, distribution networks should be restored by reconfiguring switches automatically with smart grid technologies. Although there have been several restoration algorithms developed to find the new network configuration, they might fail to restore the whole network if the network were critically damaged. The networks design has to guarantee that it is restorable under any possible failure for secure power delivery, but it is a computationally hard task to examine all possible failures in a large-scale network with complex electrical constraints. This paper proposes a novel method to find all the critical (unrestorable) line cuts with great efficiency to verify the network design. The proposed method first runs a fast screening algorithm based on hitting set enumeration; the algorithm selects suspicious cuts without naively examining all possible cuts. Next, unrestorable cuts are identified from the suspicious ones with another algorithm, which strictly tests the restorability of the network under each suspicious cut without redundantly repeating heavy power flow calculations. Thorough experiments on two distribution networks reveal that the proposed method can find thousands of unrestorable cuts from the trillions of possible cuts in a large 432-Bus network with no significant false negatives.

Evaluation of Annual Energy Loss Reduction Based on Reconfiguration Scheduling

In distribution network management, switch reconfiguration is an important tool for reducing energy loss. Recently, a variety of annual reconfiguration planning methods considering energy loss have been studied. However, no conventional methods address the reconfiguration periods in fine granularity. Practically, switch durability does not support high-frequency switching. Therefore, this paper proposes a new optimization method for annual reconfiguration scheduling. This method determines switch configurations and their reconfiguration periods with a constraint on the permissible reconfiguration times. In addition, this paper reveals the annual energy loss reduction effect of this optimization. Our method is based on partial network optimization with exhaustive enumeration of all feasible configurations. Experiments were conducted using a standard Japanese distribution network model with 468 switches. The results show that optimizing the reconfiguration periods reduces energy loss by up to 2.1 times, relative to that in a simulated conventional operation, which considers reconfiguration at equal intervals. We believe that this is the first quantitative report to address the difference between optimal reconfiguration scheduling and conventional reconfiguration.

Using \pi DDs for Nearest Neighbor Optimization of Quantum Circuits

Recent accomplishments in the development of quantum circuits motivated research in Computer-Aided Design for quantum circuits. Here, how to consider physical constraints in general and so-called nearest neighbor constraints in particular is an objective of recent developments. Re-ordering the given qubits in a circuit provides thereby a common strategy in order to reduce the corresponding costs. But since this leads to a significant complexity, existing solutions either worked towards a single order only (and, hence, exclude better options) or suffer from high runtimes when considering all possible options. In this work, we provide an alternative which utilizes so-called \(\pi \)DDs for this purpose. They allow for the efficient representation and manipulation of sets of permutations and, hence, provide the ideal data-structure for the considered problem. Experimental evaluations confirm that, by utilizing \(\pi \)DDs, optimal or almost optimal results can be generated in a fraction of the time needed by exact solutions.

Summarizing a document by trimming the discourse tree

Recent studies on extractive text summarization formulate it as a combinatorial optimization problem, extracting the optimal subset from a set of the textual units that maximizes an objective function without violating the length constraint. Although these methods successfully improve automatic evaluation scores, they do not consider the discourse structure in the source document. Thus, summaries generated by these methods may lack logical coherence. In previous work, we proposed a method that exploits a discourse tree structure to produce coherent summaries. By transforming a traditional discourse tree, namely a rhetorical structure theory-based discourse tree (RST-DT), into a dependency-based discourse tree (DEP-DT), we formulated the summarization procedure as a Tree Knapsack Problem whose tree corresponds to the DEP-DT. This paper extends the work with a detailed discussion of the approach together with a novel efficient dynamic programming algorithm for solving the Tree Knapsack Problem. Experiments show that our method not only achieved the highest score in both automatic and human evaluation, but also obtained good performance in terms of the linguistic qualities of the summaries.

On the Size of the Zero-Suppressed Binary Decision Diagram that Represents All the Subtrees in a Tree

This paper presents a method of constructing a ZDD that represents all connected subtrees in the given tree and analyzes the size of the resulting ZDD. We show that the size of the ZDD is bounded by \(O(nh)\) for a tree with \(n\)-nodes and \(h\)-height. Furthermore, by properly ordering the ZDD variables, we can further reduce the size to \(O(n\log n)\), which is surprisingly small compared to represent at most \(O(2^n)\) subtrees.

A Query-Focused Summarization Method that Guarantees the Inclusion of Query Words

Query-focused summarization is a variant of automatic summarization that uses a given query to generate summaries. It is used, for example, to generate search result snippets. Although conventional score-based methods tend to generate summaries that contain words in the given query, they offer no guarantees of that. We show that such guarantees can be formulated as an optimization problem where one constraint is that the generated summary contains at least some of the words in the query. We also show that the formulated optimization problem can be efficiently solved by using Lagrangian relaxation.

Enumerating Eulerian Trails via Hamiltonian Path Enumeration

Given an undirected graph G, we consider enumerating all Eulerian trails, that is, walks containing each of the edges in G just once. We consider achieving it with the enumeration of Hamiltonian paths with the zero-suppressed decision diagram (ZDD), a data structure that can efficiently store a family of sets satisfying given conditions. First we compute the line graph L(G), the graph representing adjacency of the edges in G. We also formulated the condition when a Hamiltonian path in L(G) corresponds to an Eulerian trail in G because every trail in G corresponds to a path in L(G) but the converse is not true. Then we enumerate all Hamiltonian paths in L(G) satisfying the condition with ZDD by representing them as their sets of edges.

Sub-sentence extraction based on combinatorial optimization

This paper describes the prospect of word extraction for text summarization based on combinatorial optimization. Instead of the commonly used sentence-based approach, word-based approaches are preferable if highly-compressed summarizations are required. However, naively applying conventional methods for word extraction yields excessively fragmented summaries. We avoid this by restricting the number of selected fragments from each sentence to at most one when formulating the maximum coverage problem. Consequently, the method only choose sub-sentences as fragments. Experiments show that our method matches the ROUGE scores of state-of-the-art systems without requiring any training or special parameters.