arXiv subjects

Eranda Çela

Publications and source records attributed to Eranda Çela.

6 recordsLinked to original sources

arXiv2023

Integrating multiple sources of ordinal information in portfolio optimization

Active portfolio management tries to incorporate any source of meaningful information into the asset selection process. In this contribution we consider qualitative views specified as total orders of the expected asset returns and discuss two different approaches for incorporating this input in a mean-variance portfolio optimization model. In the robust optimization approach we first compute a posterior expectation of asset returns for every given total order by an extension of the Black-Litterman (BL) framework. Then these expected asset returns are considered as possible input scenarios for robust optimization variants of the mean-variance portfolio model (max-min robustness, min regret robustness and soft robustness). In the order aggregation approach rules from social choice theory (Borda, Footrule, Copeland, Best-of-k and MC4) are used to aggregate the total order in a single ``consensus total order''. Then expected asset returns are computed for this ``consensus total order'' by the extended BL framework mentioned above. Finally, these expectations are used as an input of the classical mean-variance optimization. Using data from EUROSTOXX 50 and S&P 100 we empirically compare the success of the two approaches in the context of portfolio performance analysis and observe that in general aggregating orders by social choice methods outperforms robust optimization based methods for both data sets.

q-fin.PM↗

arXiv2023

A linear time algorithm for linearizing quadratic and higher-order shortest path problems

An instance of the NP-hard Quadratic Shortest Path Problem (QSPP) is called linearizable iff it is equivalent to an instance of the classic Shortest Path Problem (SPP) on the same input digraph. The linearization problem for the QSPP (LinQSPP) decides whether a given QSPP instance is linearizable and determines the corresponding SPP instance in the positive case. We provide a novel linear time algorithm for the LinQSPP on acyclic digraphs which runs considerably faster than the previously best algorithm. The algorithm is based on a new insight revealing that the linearizability of the QSPP for acyclic digraphs can be seen as a local property. Our approach extends to the more general higher-order shortest path problem.

cs.DS↗

arXiv2023

Recognising permuted Demidenko matrices

We solve the recognition problem (RP) for the class of Demidenko matrices. Our result closes a remarkable gap in the recognition of specially structured matrices. Indeed, the recognition of permuted Demidenko matrices is a longstanding open problem, in contrast to the effciently solved RP for important subclasses of Demidenko matrices such as the Kalmanson matrices, the Supnick matrices, the Monge matrices and the Anti-Robinson matrices. The recognition of the permuted Demidenko matrices is relevant in the context of hard combinatorial optimization problems which become tractable if the input is a Demidenko matrix. Demidenko matrices were introduced by Demidenko in 1976, when he proved that the Travelling Salesman Problem (TSP) is polynomially solvable if the symmetric distance matrix fulfills certain combinatorial conditions, nowadays known as the Demidenko conditions. In the context of the TSP the recognition problem consists in deciding whether there is a renumbering of the cities such that the correspondingly renumbered distance matrix fulfills the Demidenko conditions, thus resulting in a polynomially solvable special case of the TSP. We show that such a renumbering of n cities can be found in $O(n^4)$ time, if it exists.

cs.DM↗

arXiv2022

On $k$-Bend and Monotonic $\ell$-Bend Edge Intersection Graphs of Paths on a Grid

If a graph $G$ can be represented by means of paths on a grid, such that each vertex of $G$ corresponds to one path on the grid and two vertices of $G$ are adjacent if and only if the corresponding paths share a grid edge, then this graph is called EPG and the representation is called EPG representation. A $k$-bend EPG representation is an EPG representation in which each path has at most $k$ bends. The class of all graphs that have a $k$-bend EPG representation is denoted by $B_k$. $B_\ell^m$ is the class of all graphs that have a monotonic $\ell$-bend EPG representation, i.e. an $\ell$-bend EPG representation, where each path is ascending in both columns and rows. It is trivial that $B^m_k\subseteq B_k$ for all $k$. Moreover, it is known that $B^m_k\subsetneqq B_k$, for $k=1$. By investigating the $B_k$-membership and the $B^m_k$-membership of complete bipartite graphs we prove that the inclusion is also proper for $k\in \{2,3,5\}$ and for $k\geqslant 7$. In particular, we derive necessary conditions for this membership that have to be fulfilled by $m$, $n$ and $k$, where $m$ and $n$ are the number of vertices on the two partition classes of the bipartite graph. We conjecture that $B_{k}^{m} \subsetneqq B_{k}$ holds also for $k\in \{4,6\}$. Furthermore, we show that $B_k \not\subseteq B_{2k-9}^m$ holds for all $k\geqslant 5$. This implies that restricting the shape of the paths can lead to a significant increase of the number of bends needed in an EPG representation. So far no bounds on the amount of that increase were known. We prove that $B_1 \subseteq B_3^m$ holds, providing the first result of this kind.

math.CO↗

arXiv2014

Well-solvable cases of the QAP with block-structured matrices

We investigate special cases of the quadratic assignment problem (QAP) where one of the two underlying matrices carries a simple block structure. For the special case where the second underlying matrix is a monotone anti-Monge matrix, we derive a polynomial time result for a certain class of cut problems. For the special case where the second underlying matrix is a product matrix, we identify two sets of conditions on the block structure that make this QAP polynomially solvable respectively NP-hard.

math.OC↗

arXiv2011

The Wiener maximum quadratic assignment problem

We investigate a special case of the maximum quadratic assignment problem where one matrix is a product matrix and the other matrix is the distance matrix of a one-dimensional point set. We show that this special case, which we call the Wiener maximum quadratic assignment problem, is NP-hard in the ordinary sense and solvable in pseudo-polynomial time. Our approach also yields a polynomial time solution for the following problem from chemical graph theory: Find a tree that maximizes the Wiener index among all trees with a prescribed degree sequence. This settles an open problem from the literature.

math.OC↗