5281/zenodo.19812700 In engineering, management, and analytical systems, decision making often occursunder structural conflict—when input data are incomplete, internally contradictory, ordistorted. Classical probabilistic approaches treat such situations as a problem of aggregatingindependent factors. When signals conflict, these methods typically yield diffuse,poorly interpretable, and unstable results.This paper introduces an approach that represents the decision process as a coherentlandscape formed by interacting factors. In this model, probability is not an estimate of ahypothesis’s truth but rather a distribution of attention among alternatives, determinedby the degree of internal consistency of the corresponding factor configurations. In otherwords, the question shifts from “how true is an alternative?” to “how structurally does itcapture the attention of the decision-making system or agent?”Crucially, “attention” here is not a metaphor but an operationally defined quantity.The distribution P(s) is normalized,Ps P(s) = 1, and can be interpreted as an allocationof computational resources, confidence, or processing priorities among alternatives. Themodel thus provides an explicit mechanism for redistributing a limited computationalbudget toward the most coherent factor configurations.We introduce the concept of a Heuristic Decision Field (HDF), where each alternativecorresponds to a point in decision space, and its “height” is determined by the coherenceof the input signals. To quantitatively describe the landscape structure, we use entropy,peak sharpness, and a decision threshold layer.This approach does not claim to be a complete theory of decision making. Rather,it offers a general conceptual framework that reinterprets choice under conflicting factorsand can serve as a foundation for further research in multi-criteria selection, analysisof contradictory requirements, and engineering methods for conflict resolution, includingproblems akin to TRIZ.Keywords: decision making, structural conflict, coherent landscape, heuristic decisionfield, multi-criteria choice, TRIZ, attention, machine learning.
Sergey Dzhumaev (Sun,) studied this question.