This is the integration paper of the NSD/MOEH programme. Over the course of 27+ working papers, the programme has accumulated a set of structurally distinct frameworks: the Dynamic Reconnaissance-Flattening (DRF) five-stage pipeline; the Multi-Chain Multi-Tentacle (MCMT) optimisation engine; the Multi-Chain Collapse-Centred Optimisation (MCCO) framework; the path-functional optimisation framework FOG-4; and the seven-box NP Landscape Classification taxonomy. These have been developed in parallel as separate frameworks addressing different abstraction levels of the same operational problem: how to optimise on NP-Hard landscapes under the strong-Copenhagen modelling commitment. DE-5 integrates these components into a single end-to-end architecture: the integrated pipeline Π = DRF + MCMT. The composition operates as follows. A raw NP-Hard problem instance enters at the input. DRF stages 1 (reconnaissance) and 2 (flattening) classify the problem via the NP Landscape diagnostic and reduce it to a tractable form. MCMT, augmented with its multi-chain multi-tentacle structure, serves as the optimisation engine at DRF stages 3 (perspective) and 4 (diagnosis): each chain is a candidate solution trajectory; each set of tentacles is a peripheral factor scout system; the shared factor pool aggregates information across chains. DRF stage 5 (tool adaptation) postprocesses the MCMT output and produces the final optimised action. The paper's contributions are theoretical. We define the composition map Π = DRF + MCMT at the category-theoretic level, prove a single composition theorem (full proof) establishing correctness preservation, complexity bound composition (additive in component complexities, not multiplicative), and information-theoretic property preservation. We characterise the cross-box structural applicability across five NP-Hard problem classes (Multidimensional Knapsack, Graph Colouring, Vehicle Routing, Quadratic Assignment, Sequential Decision); Box 3 (SAT) is excluded by structural design. We cite the measured performance results from prior programme papers (MCCO v3's 3.56× speedup, DRF v7's per-box measurements, Spacetime DE's reconnaissance benchmarks) without generating new empirical claims. We close with the programme-level synthesis statement: the pipeline Π is the programme's articulation of how to act on the real fog landscape. This is a programme-internal Working Paper Version 1. All performance claims in this paper are either (i) theoretical (complexity bounds, information-theoretic properties, composition correctness) or (ii) cited from prior programme papers where they were measured. No new empirical claims are made in this paper. Empirical validation of the composed pipeline is the subject of a separate companion paper that will be developed when the full pipeline implementation infrastructure is in place. The paper bridges to FOG-4 v2 (10.5281/zenodo.20152165) as the discrete-time, finite-component realisation of the continuous-time path-functional framework; to DE-4 / MCMT (10.5281/zenodo.20178724) as the optimisation engine at DRF stages 3-4; to MCCO Standalone (10.5281/zenodo.20104942) as the multi-chain CPTP foundation; to NP Landscape Classification (10.5281/zenodo.19653758) as the diagnostic that drives DRF stage 1; and to DRF (10.5281/zenodo.19600597) as the five-stage pipeline structure.
H Y Rao (Thu,) studied this question.