Consideration was given to optimization of the queue control strategy in the MlGl1l queuing system where decision about continuing or stopping admission of customers is made at the service completion instants of each customer in compliance with the distribution on the set of decisions depending on the number of customers remaining in the system. The mean specific income in the stationary mode was used as the efficiency criterion and the set of permissible strategies coincided with the set of homogeneous randomised Markov strategies. It was proved that if there exists an optimal strategy, then it is degenerate and threshold with one point on control switching, that is, if the number of customers in the system exceeds a certain level, then admission of customers must be stooped or, otherwise, it must be continued.
We consider the problem of stochastic linear regulator over an infinite time horizon with superexponentially stable matrix in the equation of state dynamics. The form of the optimal control based on the criterion taking into account the information about the parameters of disturbances and the matrix stability rate was determined. The results obtained were used to analyze the model of a system with extremely impatient agents where the objective functional includes discounting by the asymptotically unbounded rate.
We find optimal (from the insurer’s point of view) strategies for insurance and reinsurance
in a controllable Cramer–Lundberg risk process that describes the capital dynamics of
an insurance company over an infinite time interval. As the optimality criterion being minimized,
we use the stationary variation coefficient, taking into account additional constraints
on residual risks for both insurers and reinsurer. We establish that it is best to use stop-loss
reinsurance with an upper limit and insurance which is a combination of a stop-loss strategy
and deductible. Equations that define optimal strategies parameters are derived.
We present the basic properties of the a new pattern analysis method in parallel coordinates; results of the method do not depend on the ordering of data in the original sample of objects being analyzed. We prove that clusters obtained with this method do not overlap. We also show the possibility of representing objects of one cluster in the form of monotonically increasing/decreasing functions.
The prototype of the isolated words recognition software based on the phonetic decoding method with the Kullback-Leibler divergence is presented. The architecture and basic algorithms of the software are described. Finally, an example of application to the problem of isolated words recognition is provided.
This paper considers a voting problem in which the individual preferences of electors are defined by the ranked lists of candidates. For single-winner elections, we apply the criterion of weak positional dominance (WPD, PD), which is closely related to the positional scoring rules. Also we formulate the criterion of weak mutual majority (WMM), which is stronger than the majority criterion but weaker than the criterion of mutual majority (MM). Then we construct two modifications for the median voting rule that satisfy the Condorcet loser criterion. As shown below, WPD and WMM are satisfied for the first modification while PD and MM for the second modification. We prove that there is no rule satisfying WPD and MM simultaneously. Finally, we check a list of 37 criteria for the constructed rules.
An axiomatics of power indices in voting with quota was proposed. It relies on the additivity and dictator axioms. Established was an important property that the player’s power index is representable as the sum of contributions of the coalitions in which it is a pivot member. The coalition contributions are independent of the players’ weights or the quota. The general theorem of power index representation and the theorem of representation for a power index of anonymous players were formulated and proved.
We show the results of a statistical study of the complexity of the asymmetric traveling salesman problem (ATSP) obtained by processing a specially generated pool of matrices. We show that the normal distribution can serve as an approximation to the distribution of the logarithm of complexity for a fixed problem dimension. We construct a family of probability distributions that represent satisfactory approximations of the complexity distribution with a dimension of the cost matrix from 20 to 49. Our main objective is to make probabilistic predictions of the complexity of individual problems for larger values of the dimension of the cost matrix. We propose a representation of the complexity distribution that makes it possible to predict the complexity. We formulate the unification hypothesis and show directions for further study, in particular proposals on the task of clustering “complex” and “simple” ATSP problems and proposals on the task of directly predicting the complexity of a specific problem instance based on the initial cost matrix.
Consideration was given to a special problem of controlling a formation of mobile agents, that of uniform deployment of several identical agents on a segment of the straight line. For the case of agents obeying the first-order dynamic model, this problem seems to be first formulated in 1997 by I.A. Wagner and A.M. Bruckstein as “row straightening.” In the present paper, the straightening algorithm was generalized to a more interesting case where the agent dynamics obeys second-order differential equations or, stated differently, it is the agent’s acceleration (or the force applied to it) that is the control.
In Russia, chain stores have achieved considerable market power. In this work, we combine a Dixit–Stiglitz industry model with a monopolistic retailer in order to address the following questions: does the retailer always impair prices, variety of goods, and ultimately welfare? Which market structure is worse: Nash or Stackelberg? What should be the public policy in this area?
In this work, we study the optimal risk sharing problem for an insurer between himself and a reinsurer in a dynamical insurance model known as the Kramer–Lundberg risk process, which, unlike known models, models not per claim reinsurance but rather periodic reinsurance of damages over a given time interval. Here we take into account a natural upper bound on the risk taken by the reinsurer. We solve optimal control problems on an infinite time interval for mean-variance optimality criteria: a linear utility functional and a stationary variation coefficient. We show that optimal reinsurance belongs to the class of total risk reinsurances. We establish that the most profitable reinsurance is the stop-loss reinsurance with an upper limit. We find equations for the values of parameters in optimal reinsurance strategies.
In order to solve robust PageRank problem a saddle-point Mirror Descent algorithm for solving convex-concave optimization problems is enhanced and studied. The algorithm is based on two proxy functions, which use specificities of value sets to be optimized on (min-max search). In robust PageRank case the ones are entropy-like function and square of Euclidean norm. The saddle-point Mirror Descent algorithm application to robust PageRank leads to concrete complexity results, which are being discussed alongside with illustrative numerical example.
The paper is concerned with scheduling the two-way traffic between two stations connected by a single-track railway with a siding. It is shown that if, for each station, the order in which trains leave this station is known or can be found, then for various objective functions an optimal schedule can be constructed in polynomial time using the method of dynamic programming. Based on this result, the paper also presents a polynomial-time algorithm minimising the weighted number of late trains.
A weakening of the property covariance named self-covariance is defined. Sels-covariant solutions are positively homogenous and satisfy a "restricted" translation covariance so that feasible shifts are only the solution vectors and their multipliers. A description of all non-empty, single-valued,efficient,anonytmous, weakly and self-covariant solutions inn the class of two-person games is given. As demonstrated below, among them there exist just three solutions admitting consistent extensions in the Davis--Maschler sense. They are the equal share solution, the stansdard solution, and the constrained ergalitarian solution for super-additive two-person games. For the third solution mentioned, characterizations of some consistent extensions to the class of all TU games are given.
This work is devoted to the problems of information transmission with frequency shift keying and fast frequency hopping in special channels where the signal/noise ratio is low, and a high energy interfering signal is present. We propose a demodulation algorithm that is significantly more stable to the influence of a powerful interfering signal as compared to other known algorithms. Under these conditions, we show a statistical criterion that lets one significantly reduce error probability on the demodulator’s output. For the chosen criterion we prove several lemmas that let us speed up the demodulation algorithm. Computer modeling results show that the proposed demodulation algorithm has better correcting ability under a powerful interfering signal than previously known ones.