|
|
 Выпуск 70
- Bocharov P., Goryashko A. Suboptimal solutions of antagonistic partition games
The paper examines (n, m)-partition games in order to develop tractable method of suboptimal decision of the resource allocation games such as colonel Blotto game or colonel Lotto game. The main goal is to develop tractable method for building suboptimal solution in mixed strategies for these games without solving the relevant optimization problem. The foundation of proposed method lies in the specific combinatorial properties of the (n, m)-partition games. It turned out that if for all game strategies the values of the balance and peculiar resource have the values from specific range it could be sufficient to get of suboptimal decisions of the games mentioned above. The proposed methods are based on both the analytical and numerical results, analytical partitions properties and numerical simulation results. The numerical simulation for the partition games (120, 6) and (100, 10) demonstrated that one could design of the epsilon-optimal decision where epsilon <= 0.02. The support set of these decisions contain no more two hundred pure strategies and decisions complexity equals const x m^2 for considered games. Results of the numerical simulation provide reasons to suppose that our approach is quite competitive with epsilon-optimal solution. The simplicity of our suboptimal solution method could be advantage in the behavioral game theory.
- Bakaev M., Razumnikova O. Defining complexity for visual-spatial memory tasks and human operator’s throughput
In this paper we study effectiveness and efficiency of HCI tasks that involve visual-spatial memory. Based on the Fitts’ law model that is widely used to model human motor reactions and inspired by Gestalt principles of human perception, we proposed formulation for index of difficulty for these tasks. The index includes the number of memorized information chunks, the size of vocabulary for the displayed objects (representing the diversity of elements), and the size of the working area, where the visual search and memorization are performed. We performed experiment with 88 subjects and the analysis of the collected data suggested that the information chunks were not elemental objects (squares), but figures composed based on range 1 von Neumann neighborhood known from cell automata. Based on the proposed index of visual-spatial memorization difficulty, we calculated throughput, which in our experiment was 3.75 bit/s (4.07 bit/s for male subjects and 3.61 bit/s for female ones).
- Melentiev V. On topological fault-tolerance of scalable computing systems
Problems of the analysis of topological fault tolerance of the scalable computing system and ensuring its sustainability to fault of the given multiplicity are considered. The measure of topological fault tolerance is offered, which connects the computing system topology with its potential parallelism for the given fault multiplicity. The relationship between the functions of topological scalability and topological fault tolerance is defined. The dependence of the minimum of a topological fault tolerance by the girth of the system graph is shown. Model of parallel computings, and functions of the topological fault tolerance and scalability are adapted to the existence of unique nodes in information topology of the solved task. A method for configuring fault-tolerant subsystems for a deficient topological fault tolerance of a computing system is proposed, while providing the preassigned fault multiplicity for the solved task is achieved by duplicating subsystems which are configured for less, than the preassigned, fault multiplicity.
- Rostova E., Geraskin M. The optimal cost function of the firm to prevent industrial risks
The problem of determining the cost function of the firm to prevent industrial risks (voluntary risk costs), the optimal by the criterion of total costs. For the exponential function of production costs and exponentially decreasing functions of the industrial damage is determined as a function of voluntary risk costs. The analysis of influence of parameters of the function of industrial damage to the area of existence of the solution of the profit maximization problem of the firm under optimal functions of voluntary risk costs. Produced by testing the theoretical results on numerical examples, provides a graphic illustration of simulation results. Analyzed the impact of the market price of commodities, production costs of firms in the voluntary risk costs. The resulting model allows for each case, describing the production process of the enterprise, to establish the boundaries of market prices and parameters of the cost function, in which the management of industrial risk is critical for the existence of the company.
- Losev A., Levshinsky V. The thermometry data mining in the diagnostics of mammary glands
The paper deals with the microwave thermometry data and proposes a method for forming informative features, based on qualitative descriptions of medical knowledge about the behavior of the temperature fields of mammary glands. There are some already known qualitative characteristics of breast cancer, which were form a basis for quantitative features, e.g. feature ‘a large thermal asymmetry between mammary glands’ can be described by temperature differences between corresponding points of right and left mammary glands. If the value of such difference is sufficiently large, then it may be an indication of pathology. After preprocessing, which involves weighting and cleaning, informative features may be applied in various classification algorithms, such as logistic regression, which yields about seventy percents of accuracy on a test sample, or the more complex ones that yield better accuracy: neural networks, genetic algorithm, and fuzzy classification. The significance of proposed features consists of the fact that they were formed from qualitative characteristics and each of them has a qualitative description, therefore they are of interest for further study and can be applied in diagnosis-advisory systems.
- Kuznetsov A. Organization of an agents' formation through a cellular automaton
The article deals with the algorithm for a distributed organization of the agents' formation defined by a graph and the numerical simulation of such algorithm. Agents move through terrain with many random obstacles (``random landscape''). At first, we describe the continuous statement of the two-criteria minimization problem. The first criterion is the agent's route time. The second criterion is the closeness of agents' formation to the desired one. Next, we introduce a cellular automaton simulating the movement of agents for obtaining quasi-optimal solutions of the problem. The cellular automaton has one-dimensional and two-dimensional representations. Agents use reflexion to predict the motion of other agents. Then we compare obtained solutions with optimal ones for different types of random landscapes via numerical experiment. At this point, we obtain the empirical distribution for the time of an agent's exit to finish point. Finally, we find the relation between a type of random landscape through which agents move and the quality of agents' formation.
|
|
