Science References

Each entry is work the papers name — lineage the method rests on, and the neighbors they position against. The comparisons and measured boundaries live in the papers; the publications page carries them.

Core Concepts & Proof of Concept

[1] Autonomous Configuration of Network Parameters in Operating Systems using Evolutionary Algorithms

This paper provided the foundational proof-of-concept work that inspired this project, demonstrating the feasibility of applying evolutionary algorithms to configure network parameters.

Algorithmic Optimization & Parallelization

[2] A unified view of parallel multi-objective evolutionary algorithms

This work underpins the project's approach to accelerating metaheuristics through parallelization, offering a unified perspective on parallel multi-objective evolutionary algorithms.

Probe-Block Design

[3] Stochastic Designs in Event-Related fMRI

Friston, Zarahn, Josephs, Henson and Dale, 1999. The block-design principle: sustained stimulation blocks alternating with rest, because brief events are smeared by the substrate's response. A push/pause probe pair is this design — a perturbation block long enough to survive the substrate's response latency, followed by a rest block that captures recovery.

[4] Statistics for Experimenters: An Introduction to Design, Data Analysis, and Model Building

Box, Hunter and Hunter, 1978 (Wiley). The ABA structure — stimulus, rest, stimulus — is standard experimental practice; probe pairs use it so every reading can be compared against its own baseline.

Signal Detection

[5] Adaptive detection mode with threshold control as a function of spatially sampled clutter-level estimates

Finn and Johnson, 1968 (RCA Review). Constant-false-alarm-rate detection, established in radar signal processing. The step-change detector that turns probe responses into coupling verdicts belongs to this family; what is new in godon is the protocol around it — who probes, who holds still — not the detector.

Sampling Rules for the Walk

[6] On the Experimental Attainment of Optimum Conditions

Box and Wilson, 1951. Response-surface methodology — sample where the response surface says information pays. The walk's sampling rule is close kin.

[7] Efficient Global Optimization of Expensive Black-Box Functions

Jones, Schonlau and Welch, 1998. Model-based design of experiments: sample where uncertainty or disagreement is highest. The walk applies the same rule inside a live substrate it does not own and cannot pause.

Composing Measured Blocks

[8] Nonlinear Problems in Random Theory

Wiener, 1958 (The Technology Press and Wiley). The origin of modeling a nonlinear system as a composition of measured blocks.

[9] Block-oriented Nonlinear System Identification

Giri and Bai (eds.), 2010 (Springer). The Wiener–Hammerstein lineage in textbook form: identify each block separately, then compose. godon's composition checks follow this material.

[10] Analysis and Simulation of a Fair Queueing Algorithm

Demers, Keshav and Shenker, 1990. Fair queueing and the max-min fairness lineage behind it.

[11] Efficient Fair Queuing Using Deficit Round-Robin

Shreedhar and Varghese, 1996. Deficit round-robin, the efficient instantiation of fair sharing. When several probing runs contend for leases, the schedule that shares turns among them descends from this work — applied to lease acquires instead of packets.

Network Tomography

[12] Network tomography: estimating source-destination traffic intensities from link data

Vardi, 1996. Inferring internal structure from external measurements.

[13] Internet Tomography

Coates, Hero, Nowak and Yu, 2002. Active probing that selects its measurements adaptively. Both share godon's active-probing philosophy; the difference is that the probers are the coupled nodes themselves, not a central observer.

Causal Framing

[14] Causality: Models, Reasoning, and Inference

Pearl, 2009. The what-if framework for interventions — confounding, do-calculus, counterfactuals. Separating coupling from confounded co-movement is the detection work's core problem, and it is this problem statement carried into a new setting: the experimenters are the coupled nodes themselves. The papers run parallel to the causal hierarchy, not on it.

Cybernetics Roots

[15] Cybernetics: Or Control and Communication in the Animal and the Machine

Wiener, 1948. Perturbing a coupled system and observing its response to discover hidden structure is, at its root, cybernetics — the study of control and communication in coupled systems.

[16] An Introduction to Cybernetics

Ashby, 1956. The law of requisite variety: a controller must model the complexity of the environment it acts on. The connectome is that model, measured empirically rather than specified analytically.

Named Neighbors

The papers position godon against these; the entries say what each is, the papers measure where each stops.

[17] System Identification: Theory for the User

Ljung, 1999 (Prentice Hall). System identification: the mature neighbor for the quiet single-owner case — one owner, a model structure specified a priori, a system that can be paused.

[18] Detecting Causality in Complex Ecosystems

Sugihara, May, Ye, Hsieh, Deyle and Fogarty, 2012. Convergent cross-mapping: coupling inferred from time series. The detection paper measures where passive methods stop separating coupling from co-movement while the system itself is exploring.

[19] Detecting and quantifying causal associations in large nonlinear time series datasets

Runge, Nowack, Kretschmer, Flaxman and Sejdinovic, 2019. Causal association quantification for nonlinear time series — passive, like convergent cross-mapping, and measured against the same boundary in the detection paper.

[20] Neural Relational Inference for Interacting Systems

Kipf, Fetaya, Wang, Welling and Zemel, 2018 (ICML). Interaction graphs learned from trajectories; operates passively — the contrast is measured in the detection paper.