Deterministic -server conjecture
The -server conjecture states that a deterministic online algorithm can achieve competitive ratio on every metric space. We prove the conjecture. Specifically, we show that the work function algorithm satisfies it.
Our proof uses a natural algebraic representation of the work function as a matrix, which encodes all feasible paths to reach a configuration. In this representation, the minimum and addition operations arising in the definition of optimal costs correspond to addition and multiplication of formal expressions, and each work function value corresponds to the determinant of columns of the matrix. A request arrival updates the representation via a change of basis and row replacement. The amortized analysis is based on a potential function defined in terms of a larger matrix whose coordinates are pairs of coordinates of the original matrix representation.
- Result
- Proved(see note)
- Status
- Resolved
- AI contribution
- —
- Method
- Argument
- Field
- Online algorithms
- Posed by
- Manasse, McGeoch, and Sleator
- Year posed
- 1988
- Years open
- 38y
- Solved
- 2026-09-14
- Model
- GPT 6 Astra
- Vendor
- OpenAI
- Collaborators
- Christian Coester, Elias Koutsoupias, Marek Zbysiński
- Verification
- Unreviewed
- Publication
- Preprint
- Significance
- —
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
The paper proves that the work function algorithm (WFA) is -competitive for the -server problem on every metric space. Specifically, for initial configuration and every finite request sequence ,Since a competitive ratio below is impossible in general, this settles the deterministic -server conjecture with the optimal ratio.
The proof represents work-function values as valuations of determinants over a formal field and constructs a potential from quadratic products of the determinant columns. Its increase at each request pays the extended cost of WFA, yielding the global bound.
This does not concern the separate randomized -server conjecture, which was previously refuted.
What the AI did
The authors first developed, without AI assistance, a potential function proving the previously unknown case on arbitrary metrics. Discussions with ChatGPT 5.5 Pro and Gemini 3.1 Pro helped them understand and symmetrize that potential. Based on this reformulation, ChatGPT 6 Astra derived an algebraic proof valid for arbitrary , thereby supplying the central generalization that proves the conjecture. The paper adapts this proof using an explicit column representation of work functions. After the authors supplied that representation, Astra further adapted the proof and assisted with drafting portions of the paper.
Verification
This is an arXiv v1 preprint claiming a proof of a longstanding major conjecture. The paper contains a complete mathematical argument, but no independent expert verification, peer review, or formal machine-checked proof is identified in the source. The claim should therefore remain classified as unreviewed pending independent scrutiny.
Source
- PaperarXiv
Submitted by VibeGene on