Parametric pricing methods and neural network analysis of product digital twin data

Abstract

With the Russian military-industrial complex operating in mobilization mode, a systemic contradiction has been identified between the highly dynamic changes in design and technological solutions and the inertia of economic justification for pricing state defense procurement products, based on a cost-based methodology and multi-iteration documentary verification procedures. The aim of this study is to develop a scientifically tool for radically reducing the cycle time for forecast pricing of weapons, military, and special-purpose equipment while maintaining the requirements of regulatory auditability and statistical reliability of calculations. A concept for an organizational and economic digital twin of an enterprise is proposed, ensuring the end-to-end integration of engineering and economic contours through a formalized ontological model represented as a composite of "design-technology-cost" mappings. An algorithm has been developed for the automated extraction of design and technological cost predictors from the electronic structure of a product and electronic production technology, followed by aggregation of the parameters into a vector of cost factors. The key result is the synthesis of an original hybrid parametric model for estimating production labor intensity, based on adaptive weighting of interpretable cost estimation ratios and neural network approximation. The weighting coefficient is determined by a logistic function of the training sample size and representativeness, ensuring automatic adaptation of the model to the enterprise's level of digital maturity and resolving the conflict between the requirements of transparency for regulatory authorities and the need to improve predictive accuracy. The architecture of the final express price justification model is strictly isomorphic to the normative costing structure established by current legislation on state price regulation, guaranteeing the reproducibility and auditability of calculations. The obtained research results substantiate the feasibility of transitioning from a document-intensive reactive pricing model to a proactive, model-oriented paradigm for managing the cost of state defense procurement (GDO) products. The development and calibration of parametric tools using domestic data and integration into national PLM systems is seen as an element of ensuring technological sovereignty and improving the efficiency of budgetary funds by reducing transaction costs and eliminating administrative delays in the price approval process.