COMPREHENSIVE METHODOLOGY FOR MULTICRITERIA QUALITY ASSESSMENT OF INFORMATION-EDUCATIONAL RESOURCES

Abstract

The modern stage of digitalization in the educational space is characterized not only by the exponential growth of the software market but also by a fundamental paradigm shift in the creation and consumption of information-educational resources (IER). In an environment of fierce competition and market saturation, product quality transforms from a desirable characteristic into a key factor for survival and commercial success. However, the process of developing and improving IER faces a profound dialectical contradiction: developers must satisfy the constantly growing demands of consumers, who seek to maximize the economic efficiency of the software, while operating under strict constraints of time, financial, and human resources. This necessitates a delicate balance between minimizing development costs and achieving high functional suitability. The study identifies that the complexity of building a universal quality assurance system is exacerbated by the lack of unified mechanisms capable of effectively modeling quality for both mass-market products and individual custom orders. Existing international and national standards, such as ISO 9126, serve as a fundamental classification basis but remain largely framework-oriented. They describe how to define properties like reliability or maintainability but fail to set specific metric systems or normative values for the educational domain, leaving this task to the discretion of users. Furthermore, traditional hierarchical assessment methods often ignore complex non-linear interdependencies and feedback loops between conflicting quality criteria - for instance, when increasing functionality negatively impacts usability - rendering them inadequate for modeling real-world systems. To address these challenges, the paper proposes a comprehensive methodology and mathematical framework for multicriteria quality assessment of IER. The core of the proposed approach is a hybrid model that synthesizes the Analytic Hierarchy Process for structuring the problem with the Analytic Network Process. This combination allows for a systemic view of quality, where metrics do not exist in isolation but interact within a network of causal relationships, enabling the model to account for the mutual influence of criteria from different clusters, such as general quality factors and specialized characteristics. A significant innovation of the study is the integration of Fuzzy Logic into the qualimetric analysis procedures to mitigate the subjectivity inherent in expert evaluations. Recognizing that human thinking operates primarily with qualitative categories, the replacement of rigid numerical scales with fuzzy sets and linguistic variables (e.g., "low complexity," "high efficiency") is advocated. This approach effectively formalizes the vague, intuitive concepts used by methodists and pedagogues, leveling the errors associated with human judgment under uncertainty and aggregating diverse expert opinions into a single integral indicator without losing critical semantic nuances. The methodology is deeply rooted in the concept of Value-Based Software Engineering. To ensure economic rationality, the study introduces specialized utility functions, specifically the multiplicative Stone’s function and the specific shifted ideal method. These mathematical tools allow for the formalization of strict consumer and budgetary constraints, automatically filtering out project variants that do not meet critical requirements or exceed cost limits. This mechanism facilitates the identification and elimination of "gold plating" - redundant features that consume significant resources but offer low value to the end-user. Consequently, the approach enables a rational redistribution of the budget towards critical quality components, ensuring the creation of a competitive product at an acceptable price. Finally, the study outlines a strategic vector for future research: the transition from discrete expert surveys to automated systems based on Learning Analytics and Big Data. This evolution aims to verify quality through the deep analysis of real user behavior logs- such as reaction times, error patterns, and navigation paths- transforming quality assurance into a continuous, data-driven cycle of improvement.