Some possible ways of implementing aggression as one of the mechanisms for the social behavior formation in robots groups are discussed in this work. Aggression is considered as a way to resolve conflicts over resources. The features of the aggressive behavior of eusocial insects (ants) are used as basic. A reactive model of behavior was proposed. The aggressive component is integrated into the demand-emotional architecture of the animat's control system, which is presented as a hybrid neuroproduction system. Also, the question of using an aggressive component at the phenomenological level of behavior management. Imitation modeling experiments were carried out on the example of realization of domination in a group. The issue of determining the basic mechanisms for feed areas distribution, which is part of the foraging task, is also considered. It is shown that such mechanisms are domination (as the result of aggressive actions) and the animats memory. The simulation results confirm that the addition of the "aggressiveness" parameter to the control system provides a variety of animats behavior taking into account the environment state. The proposed model of aggressive behavior does not depend on the task, which is solved, and allows to manage the group in natural form.
New algorithms of patterns analysis based on methods of ordinal-fixed and ordinal-invariant pattern clustering are developed. The definition of the proposed methods as well as the evaluation of the computational complexity is given. We provide some examples that demonstrate features of these clustering procedures and explain their operation. We also formulate and prove the theorem on the interconnection of clusters obtained by the use of ordinal-invariant pattern-clustering with complete weighted digraphs. These results allow to apply graph theory for the study of properties of obtained clusters.
The article proposes a scheme of diagnosing the problems of complex systems development, based on the construction of a cognitive map of the system, and its supporting methods. It’s allows: explore different types of problems related to the interaction of active agents, the adverse effect of the environment and the structural features of the system; consider the problem in the complex and to determine priorities for problem solving, depending on their causes.
The research of queueing model with controlled semi-markov batch flow is carried out. CBSMAP-flow is a generalization of BMAP-flow. Different variants of construction for control set are presented. Income functional is constructed on the trajectories of the controlled semi-markov process. Theorems about struc-ture of income functional for the new researched queueing model is proved.
Quantitative evaluation of the effectiveness and risks of implementa-tion of information systems (IS) is essential to start the project. It is necessary to choose a reasonable performance indicators and the method of determining their dependence on the variables affecting the project. The Monte Carlo method is proposed for the analysis, and net present value of the cash flows of the project is selected as a criterion of effectiveness. We discuss external factors affecting the effectiveness, give recommendations how to determine their proba-bilities and present the project assessment model. We also propose a method to calculate the positive cash flow which occurs due to the reduction of time of works supported by IS. Impact of deviations from project schedule is taken into account. The practical case of the project analysis for the implementation of knowledge management IS is presented.
Abstract: An incentive model for an organizational system (OS) agent is considered, using a combination of the method of inverted priorities and a modified counterplan method. The resulting model motivates the agent to report the planned KPI value, which coincides with the agent's forecast and at the same time stimulates the agent to submit an adequate application for the resource. The model allows the OS center to allocate a dedicated resource to run each KPI between agents so that in addition to reporting adequate plans, agents are committed to their implementation. Restrictions are imposed on the model parameters, in which the choice of an agent equal to its own forecast is a dominant strategy (DS), provided that the agent submits the optimal (according to the method of inverted priorities) requests for the resource. It is proved that such a choice for all OS agents is an equilibrium in DS. The procedure for determining the planned KPI values for OS agents and resource allocation between agents is described. The model parameters determined by the OS center, taking into account the imposed restrictions, allow selecting the required priorities for agent motivation and distributing a motivational fund in accordance with the priorities identified by the center. An example of the application of the model for stimulating scientific laboratories for the preparation of students is given.
In the article we examine the task of planning for the alternative paths for an unmanned vehicle. This task is a key part of a bigger problem – multi agent path planning, e.g. finding the set of non-conflict paths for a coalition of vehicles. We propose a new path planning method which indirectly takes into account vehicle’s movement dynamics and guarantees the feasibly of the resultant paths. As well we elaborate on a number of modifications of proposed method. We conduct empirical study of all the introduced algorithms by running large number of the experiments simulating nap-of-the-earth flight of compact multirotor unmanned aerial vehicle in urban environment.
Abstract: We consider a three-parameter model for stimulating the agent of the organizational system (OS), which uses the modification of the counter-plan method. For the proposed model, restrictions are introduced on the parameters under which an agent's choice equalы to his own forecast and is the dominant strategy (DS) of the agent. It is proved that such a choice for all organizational system agents is an equilibrium in dominant strategy. An example of the application of the model for stimulating units for dealing with distressed assets (PRPA) in the North-West Bank of the Sberbank of the Russian Federation is given. There is a more accurate determination of the PPRA plans for the proposed model in comparison with the traditional "top-down" planning method.
Information systems implementation efficiency measurement method is proposed. It is based on entropy of characteristic parameter of business process which is supported by system. Equations are introduced, which help to estimate the rate of processes with predefined output. Example is given of the proposed method application to information system effect estimation and finding the direction of business processes improvement.
In this paper, we consider the classical statistical problem of probability density estimation based on a sample from this distribution. This problem naturally arises in many applications when one aims at investigation of a probability structure in a random process. For instance, it is possible to identify some structure in a complex system using density estimation. In this paper, a new approach to estimate a density function is proposed. This approach is based on approximation of a log-density via Fourier series with coefficients obtained by solving a system of linear equations. Analysis of theoretical properties of such an estimate is the main purpose of this work. As the main results, we prove bounds on the difference between target density and its approximation in the supremum norm and the Kullback-Leibler divergence. Obtained rates are parametric and have order 𝑂(1/√𝑁) with high probability, which is a standard rate in parametric estimation problems. The constants in the rates are obtained up to an absolute factor, which means that we investigated the dependence on all parameters. As a numerical example, we consider a problem of Cauchy density estimation.
One of the most important branches of cognitive-map-based tools development is constructing a cognitive map being an integrated model of knowledge by a group of experts. We survey the procedures for group maps construction suggested by leading theo-rists and practitioners and analyze the risks incurred by these tech-niques and their reliability. We show that a typical procedure of building a group map employs formal aggregation (averaging) of expert estimates with no analysis of experts’ points of view and no necessary reconciliation. We suggest a number of principles and routines to expert estimates’ reconciliation and clusterization, which result in a more reasonable opinion aggregation.
Under the assumption that there exists a comprehensive hierarchical classification of the sciences, the level of research impact of a scientist is defined as the hierarchical level of the field of science created or transformed by the scientist. Current indexes for scoring the research impact are critically discussed with respect to this definition. Ways to make the scoring more adequate are highlighted.
The work continues the research of constructing methods for analyzing patterns in parallel coordinates independent of the sequence of input data of the results. The basic operations on objects of ordinal-invariant pattern clusters are described. The assertion that the centroid of an ordinal-invariant pattern cluster belongs to the original cluster is proved, which allows one to estimate the intracluster object - centroid distances in the multidimensional feature space. Examples of revealing the structural similarity of objects in parallel coordinates are given. The main differences between the methods of analysis of patterns and cluster analysis are noted. The methodology of the centroid detection of the ordinal-invariant pattern- cluster is described. An algorithm for combining groups of objects based on their structural similarity, on the one hand, and minimizing intracluster distances, on the other, is proposed, which makes it possible to improve the accuracy of the final results and partially solve the problem of finding similar objects in the presence of error in the original data. The proposed algorithm uses the concept of intracluster distances “object - centroid” and satisfies the following conditions: endogenous determination of the number and composition of the desired groups of objects under study; low (relatively) computational complexity; independence of the original partition from the initial sequence of input data. The work of the proposed algorithm on classical data sets is demonstrated. The results of testing are presented and the clustering accuracy is increased.
We consider a problem of optimal portfolio selection in order to track a riskless reference portfolio. The performance of controls is evaluated according to the investor’s time-preference. We investigate the stochastic optimality of the control which minimizes the expected long-run cost, providing an asymptotically upper estimate (almost surely) on difference between objective functionals corresponding to the optimal control and any admissible control strategy.
The new approach to evaluation of scientists’ output is proposed based on aggregation of separate bibliometric indicators using the procedure of threshold aggregation. The method is illustrated on a model dataset.