Method of Optimization of Information Models Scaled in Space of Analytical Web-Systems by the Criteria of Completeness of their Topological Observation
Keywords
Abstract
The article considers the issue of optimization of information models of space-scaled analytical web systems by the criterion of completeness of their topological observability.
The previously created, with the participation of some co-authors, information technology analysis and optimization of topological observability of multilinked geographic information systems, i.e. — optimization of the system model in the form of a dichromatic graph, which allows arbitrary input variables to obtain values of state variables or current system parameters. information about the current state to synthesize the law of its management. In such a bichromatic graph, some vertices are variable and others are dependencies between them. This technology introduces the criterion of observability, which can take values from 0 to 1 and compare different models according to the level of their observability and optimize them by choosing the most observable option. This paper proposes how such a model formalization and this technology can be extended to the formalization of space-scaled analytical web systems. Then a new method of optimizing information models of such systems by the criterion of completeness of their topological observability was proposed, which allows to analyze the completeness of this observability both at each level of scaling in space where analytical operations are planned in the projected web system and for the whole systems in general. It is proposed to optimize the model by introducing additional vertices-variables and vertices-dependencies between different levels. That is, in the absence of data or dependencies at one level of the system, they can sometimes be obtained from data or dependencies at another level, and vice versa, which will increase the observability of the system as a whole. The method allows both to analyze the level of topological observability and to analyze it after optimization. As a result of such analysis, it is possible to impose software constraints on the functionality of the web system to avoid queries or dependencies that are not provided with data or dependencies from which they could be obtained due to incomplete topological observability of individual (or all) levels.
The paper demonstrates the work of the method on the example of developing an automated web system for calculating and forecasting the water balance of the Dniester basin. The application of the developed approach at the design stage of this system ensured the avoidance of queries or dependencies not provided with data. The system was developed and successfully implemented.
The developed method of optimizing information models of space-scaled analytical web systems by the criterion of completeness of their topological observability can be used in the analysis and design of systems that reflect environmental or infrastructure data distributed in space and have several levels of detail. These can be cadastral systems, environmental monitoring systems, etc.
