Evolutionary optimization: a big data perspective

Stochastic search techniques such as evolutionary algorithms (EA) are known to be better explorer of search space as compared to conventional techniques including deterministic methods. However, in the era of big data like most other search methods and learning algorithms, suitability of evolutionary algorithms is naturally questioned. Big data pose new computational challenges including very high dimensionality and sparseness of data. Evolutionary algorithms' superior exploration skills should make them promising candidates for handling optimization problems involving big data. High dimensional problems introduce added complexity to the search space. However, EAs need to be enhanced to ensure that majority of the potential winner solutions gets the chance to survive and mature. In this paper we present an evolutionary algorithm with enhanced ability to deal with the problems of high dimensionality and sparseness of data. In addition to an informed exploration of the solution space, this technique balances exploration and exploitation using a hierarchical multi-population approach. The proposed model uses informed genetic operators to introduce diversity by expanding the scope of search process at the expense of redundant less promising members of the population. Next phase of the algorithm attempts to deal with the problem of high dimensionality by ensuring broader and more exhaustive search and preventing premature death of potential solutions. To achieve this, in addition to the above exploration controlling mechanism, a multi-tier hierarchical architecture is employed, where, in separate layers, the less fit isolated individuals evolve in dynamic sub-populations that coexist alongside the original or main population. Evaluation of the proposed technique on well known benchmark problems ascertains its superior performance. The algorithm has also been successfully applied to a real world problem of financial portfolio management. Although the proposed method cannot be considered big data-ready, it is certainly a move in the right direction.

Appropriate choice of aggregation operators in fuzzy decision support systems

Fuzzy logic provides a mathematical formalism for a unified treatment of vagueness and imprecision that are ever present in decision support and expert systems in many areas. The choice of aggregation operators is crucial to the behavior of the system that is intended to mimic human decision making. This paper discusses how aggregation operators can be selected and adjusted to fit empirical data—a series of test cases. Both parametric and nonparametric regression are considered and compared. A practical application of the proposed methods to electronic implementation of clinical guidelines is presented

Monotone approximation of aggregation operators using least squares splines

The need for monotone approximation of scattered data often arises in many problems of regression, when the monotonicity is semantically important. One such domain is fuzzy set theory, where membership functions and aggregation operators are order preserving. Least squares polynomial splines provide great flexbility when modeling non-linear functions, but may fail to be monotone. Linear restrictions on spline coefficients provide necessary and sufficient conditions for spline monotonicity. The basis for splines is selected in such a way that these restrictions take an especially simple form. The resulting non-negative least squares problem can be solved by a variety of standard proven techniques. Additional interpolation requirements can also be imposed in the same framework. The method is applied to fuzzy systems, where membership functions and aggregation operators are constructed from empirical data.

Global profit split: An evolutionary approach to international income allocation

International taxation is concerned mainly with the equitable allocation of cross-border income between countries in which income-earning activities take place. Such allocation has traditionally been governed by the arm’s-length principle, which has been interpreted as requiring a comparable transactional pricing approach. This approach assumes that each member of a multinational enterprise (MNE) group is a separate entity and that the transactions between related parties can be separated and compared with arm’s-length transactions. It has, however, proved difficult to apply comparable transactional pricing to internationally integrated businesses, especially those involving intangibles and services, and formulary apportionment has been suggested as an alternative. Essentially, formulary apportionment treats the MNE group as a single economic entity. The group’s profit is allocated to members according to a formula that reflects the particular member’s contribution to the production of that profit. A rich academic literature exists which either defends or attacks this alternative approach. The OECD and national governments have rejected formulary apportionment mainly on the ground that it violates the arm’s-length principle. This article proposes a global profit split (GPS) method for allocating international income. The GPS would allocate the global profit of an integrated business to each country in accordance with the economic contributions made by components of the business located in that country. The allocation would be based on a formula that would reflect the economic factors that contribute to profit making. While the GPS draws on elements of the traditional formulary apportionment and profit split methods, it also differs from them. The author discusses in detail the key issues involved in designing the GPS. She also presents and evaluates the main policy and pragmatic justifications for the adoption of this innovative approach. The author argues that the GPS is not only theoretically and practically superior to traditional income allocation methods, but also consistent with the arm’s-length principle. On the basis of historical developments, interpretation of article 9 of the OECD model tax convention, and international tax policy considerations, the author establishes that the GPS is not a radical departure from the arm’s-length principle, but rather a natural development in its evolution. She concludes that the law of evolution ison the side of reform because the GPS would provide for a fair and effective allocation of income derived from globally integrated business activities.

How to build aggregation operators from data

This article discusses a range of regression techniques specifically tailored to building aggregation operators from empirical data. These techniques identify optimal parameters of aggregation operators from various classes (triangular norms, uninorms, copulas, ordered weighted aggregation (OWA), generalized means, and compensatory and general aggregation operators), while allowing one to preserve specific properties such as commutativity or associativity. © 2003 Wiley Periodicals, Inc.

Fitting generated aggregation operators to empirical data

This paper treats the problem of fitting general aggregation operators with unfixed number of arguments to empirical data. We discuss methods applicable to associative operators (t-norms, t-conorms, uninorms and nullnorms), means and Choquet integral based operators with respect to a universal fuzzy measure. Special attention is paid to k-order additive symmetric fuzzy measures.

Learning weights in the generalized OWA operators

This paper discusses identification of parameters of generalized ordered weighted averaging (GOWA) operators from empirical data. Similarly to ordinary OWA operators, GOWA are characterized by a vector of weights, as well as the power to which the arguments are raised. We develop optimization techniques which allow one to fit such operators to the observed data. We also generalize these methods for functional defined GOWA and generalized Choquet integral based aggregation operators.

Methods of construction of OWA operators from data

This paper investigates the problem of obtaining the weights of the ordered weighted aggregation (OWA) operators from observations. The problem is formulated as a restricted least squares and uniform approximation problems. We take full advantage of the linearity of the problem. In the former case, a well known technique of non-negative least squares is used. In a case of uniform approximation, we employ a recently developed cutting angle method of global optimisation. Both presented methods give results superior to earlier approaches, and do not require complicated nonlinear constructions. Additional restrictions, such as degree of orness of the operator, can be easily introduced

Evolutionary structure learning algorithm for Bayesian network and penalized mutual information metric

This paper formulates the problem of learning Bayesian network structures from data as determining the structure that best approximates the probability distribution indicated by the data. A new metric, Penalized Mutual Information metric, is proposed, and a evolutionary algorithm is designed to search for the best structure among alternatives. The experimental results show that this approach is reliable and promising.

Identification of general and double aggregation operators using monotone smoothing

Aggregation operators model various operations on fuzzy sets, such as conjunction, disjunction and averaging. Recently double aggregation operators have been introduced; they model multistep aggregation process. The choice of aggregation operators depends on the particular problem, and can be done by fitting the operator to empirical data. We examine fitting general aggregation operators by using a new method of monotone Lipschitz smoothing. We study various boundary conditions and constraints which determine specific types of aggregation.

Identification of general aggregation operators by Lipschitz approximation

Aggregation operators model various operations on fuzzy sets, such as conjunction, disjunction and aver aging. The choice of aggregation operators suitable for a particular problem is frequently done by fitting the parameters of the operator to the observed data. This paper examines fitting general aggregation operators by using a new method of Lipschitz approximation.