Variable neighborhood search heuristic for nonconvex portfolio optimization

Abstract

In this article we consider a portfolio optimization problem under multiple real-world constraints, such as: cardinality constraints, tracking error, active share, and turnover. We propose a heuristic based on variable neighborhood search (VNS) that effectively addresses additional constraints that introduce non-convexities. In the VNS-based heuristic, several neighborhood structures are introduced and fast local search is implemented. We develop a VNS portfolio rebalancing framework (VNS-PRF) with two rebalance strategies. Data sets provided by a financial investment firm are used to evaluate the validity and reliability of the proposed VNS-PRF. Computational experiments and different portfolio performance measures indicate that our approach is able to obtain solutions with competitive quality and can be applied on large-scale data sets.     

Does adding up of economic capital for market- and credit risk amount to conservative risk assessment?

There is a tradition in the banking industry of dividing risk into market risk and credit risk. Both categories are treated independently in the calculation of risk capital. But many financial positions depend simultaneously on both market risk and credit risk factors. In this case, an approximation of the portfolio value function separating value changes into a pure market risk plus pure credit risk component can result not only in an overestimation, but also in an underestimation of risk. We discuss this compounding effect in the context of foreign currency loans and argue that a separate calculation of economic capital for market risk and for credit risk may significantly underestimate true risk.     

RETRACTED ARTICLE: Analyzing of innovations influence on economic growth by fuzzy system

Economic growth may be developed on the basis on combination of different factors. In this investigation was analyzed the economic growth prediction based on the innovations by field of technology. Gross domestic product (GDP) was used as economic growth indicator. The method of adaptive neuro fuzzy inference system (ANFIS) was applied to the data in order to detect the influential parameters for the GDP prediction. Five inputs are considered: number of granted patents in electrical engineering, number of granted patents as instruments, number of granted patents in chemistry, number of granted patents in mechanical engineering and the number of granted patents in other fields. Results shown that the innovations in electrical engineering has the highest influence on the GDP prediction.

     

Determination of the most influential factors for number of patents prediction by adaptive neuro-fuzzy technique

Number of patents may be developed on the basis on different natural and science and technological factors. Number of patents prediction based on the different factors in many countries is analyzed in this investigation. These factors represent natural and science resources. The method of ANFIS (adaptive neuro fuzzy inference system) was applied to the data in order to select the most influential factors for the number of patents prediction. Five inputs are considered: research and development (R&D) resources, natural resources, quality of academic institutions, quality of collaboration with the private sector and quality of education. As the ANFIS output, number of patents is considered. The ANFIS process for variable selection is also implemented in order to detect the predominant factors affecting the prediction of number of patents. Results show that the R&D is the most influential factor for the number of patents prediction.     

Person-level assessment of measurement invariance

Quality & Quantity - Existing procedures for testing measurement invariance focus mainly on group-level comparisons rather than individual comparisons (i.e., the main conclusion typically...     

‘I want to ride my bicycle’: delimiting cyclist typologies

Based around fieldwork in Seville (Spain), our article provides an empirical analysis with the aim of determining whether different typologies of cyclists exist depending on the type of bicycle for urban commuting (public bicycle/private bicycle). Our findings show that users of public bicycles are predominantly male, young, with a high level of education, and basically use the public bicycle for subsistence trips due to its easy intermodality; while private bicycle riders are mainly females who regularly make nonsubsistence trips and prefer a more flexible bicycle for their daily needs.     

Spatio-temporal determinants of the structural and productive transformation of regions in Central and East European countries

This paper investigates the determinants of the structural (between-sectoral) and productive (within-sectoral) transformation of 56 NUTS2 regions in Central and East European (CEE) countries over the 2008–2014 period. Over the analysed period, a transformation of the analysed regions has taken place towards the tertiary sector and high value-added activities. The results from a spatial panel Durbin analysis reveal the existence of product-space linkages and point to a productive transformation towards high technology-intensive activities as an important driver of industrial development. Equal access to digital infrastructure in urban and rural areas and transfer of skills, knowledge and resources through the inflow of foreign investment and imports of production inputs increase the contribution of manufacturing to regional employment. The effect of profit tax reductions on the share of manufacturing employment is about five times larger than the effect of government subsidies to production.     

Factors that determine the cost-effectiveness ranking of second-best instruments for environmental regulation

This paper develops a conceptual model to analyze how specific factors affect the compliance costs of three suboptimal policy instruments, when compared to the optimal ambient permit system (APS) benchmark. The model considers a non-uniformly mixed pollutant and explicitly incorporates the following factors: number of polluting sources; size, in terms of emissions, of each process; marginal abatement costs for each process; effluent concentrations; the transfer coefficient that relates emissions to environmental quality at the receptor; and the desired environmental quality target. APS is compared to a suboptimal emission permit system (EPS), and two Command and Control (CAC) policies—equal percentage reduction (PER) and a uniform effluent concentration standard (STD). The results show the importance of the different factors and their interactions in determining each policy instrument’s cost-effectiveness ranking. Surprisingly, EPS performs well within the usual values of these factors and in specific cases STD and PER also perform similarly to APS.      

Natural estimation of variances in a general finite discrete spectrum linear regression model

The method of “natural” estimation of variances in a general (orthogonal or nonorthogonal) finite discrete spectrum linear regression model of time series is suggested. Using geometrical language of the theory of projectors a form and properties of the estimators are investigated. Obtained results show that in describing the first and second moment properties of the new estimators the central role plays a matrix known in linear algebra as the Schur complement. Illustrative examples with particular regressors demonstrate direct applications of the results.     

Mobile Dwelling Unit Developed by a Group of Mechanical Engineering Design Students

The increasing number of homeless people is one of the most burning problems of nowadays world. Society has to find the way to fight against it and provide help to those people that have no place to stay. For short-term interventions, the application of the mobile dwelling units seems to be the only possible solution. However, existing mobile dwelling units are rather limited to specific areas of application. A group of students faced this problem and decided to develop an alternative, more universal mobile dwelling unit that will fulfil a wide range of different technical and social requirements at a reasonable price. The paper describes a systematic approach to development of the mobile dwelling unit, which was performed as a part of the education process at the Faculty of Mechanical Engineering, University of Maribor. The project presented is interdisciplinary. In addition to the educational aspect that is emphasized in this paper, technical, economic and social aspects of the project have also a very important role. From educational point of view, the project was concluded successfully. Moreover, we believe our proposal for a mobile dwelling unit is both novel and suitable for general application. This revised version was published online in July 2006 with corrections to the Cover Date.