Decision Engineering Science™ A Research and Validation Programme for Governed Decisions

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

What does it take to make a consequential decision explainable, governable, and safe to execute? This manuscript sets out a twenty-question research programme for Decision Engineering Science™ (DES). It examines how decisions can be represented as persistent objects, reconstructed from evidence, assessed for quality, checked against governance and authority requirements, revalidated as conditions change, and connected to execution and outcomes. RQ-001–RQ-020 turn these ideas into testable propositions. Each research question identifies its proposed mechanism, comparison conditions, candidate measures, ground-truth requirements, falsification criteria, and threats to validity. The programme spans individual decisions, human oversight, multi-agent systems, and interoperability across architectures. The manuscript also reports a bounded prior-art audit: 266 search-result rows across 40 fixed search arms, a one-step citation chase from 19 distinct sources, and source-element mappings for all twenty questions. A supplemental full-text comparison with Decision-making Driven Design (DMDD) identifies substantive overlaps and clarifies the remaining DES research claims. The contribution is a testable research architecture, not a claim that its effects have already been proven. Decision records, provenance, quality measures, governance controls, and temporal monitoring all have antecedents. The open question is whether the specified DES combination improves decision reconstruction and governance over strong, information-equivalent alternatives. Cross-domain validation, cross-architecture validation, empirical testing, and independent replication remain to be completed. This publication provides the questions, protocols, and evidence boundaries needed to test that claim.