Pith. sign in

REVIEW 3 minor 14 references

Asymptotic optimality of dynamic first-fit packing on the half-axis

T0 review · 0 major / 3 minor · reviewed 2026-05-24 · grok-4.3

Pith's one-line read First-fit dynamic packing converges to a gap-free size-ordered configuration as arrival rate tends to infinity

desk verdict This paper gives the first proof of the 1985 conjecture on asymptotic optimality of first-fit for non-degenerate countable-support H. read the letter →

arxiv 2404.03797 v2 pith:NPSF6WYE submitted 2024-04-04 math.PR cs.DS

classification math.PRcs.DS
keywords first-fitpackingdynamicstorageallocationasymptoticoptimalityPoissonarrivalssteady-stateconvergencehalf-axismodelitemsizedistribution
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper establishes the asymptotic optimality of the first-fit algorithm for dynamic item placement on a half-axis when item sizes come from a distribution with countable support. Items arrive according to a Poisson process with rate r and have exponentially distributed service times. The key result shows that the steady-state configuration, scaled by 1/r, converges in distribution to one where items are arranged from smallest to largest with no empty space between them. This implies that the proportion of wasted space goes to zero as r becomes large. The proof covers distributions concentrated on countably many sizes that are either finite in number or increase to infinity, with finite average size.

What carries the argument

The first-fit placement rule, which assigns each item to the leftmost feasible gap, together with the convergence in distribution of the scaled configuration to the deterministic sorted no-gap limit

What would settle it

Finding a countable-support distribution H where for large r the scaled empty space remains positive with positive probability would falsify the claim

Watch

Extended reading notes

Core claim

Under the first-fit discipline, as r tends towards infinity, the steady-state packing configuration (scaled down by r) converges in distribution to the limiting packing configuration with smaller items on the left, larger items on the right, and with no gaps between. In particular, this proves asymptotic optimality of first-fit in the sense that in steady-state the empty space (scaled down by r) vanishes.

Load-bearing premise

The distribution H of item sizes is supported on a countable increasing sequence of positive reals, finite or unbounded, with finite mean

Editorial extensions

If this is right

  • The ratio of expected empty space to expected occupied space tends to zero
  • The limiting configuration has all items packed contiguously in increasing size order from the origin
  • Asymptotic optimality holds for the specified class of item size distributions H
  • The result resolves the conjecture of Coffman, Kadota, and Shepp for non-degenerate H with countable support

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The size-ordering may indicate that first-fit achieves the optimal packing density in the fluid limit
  • The proof technique might extend to distributions with continuous support though the paper restricts to countable cases
  • Other online algorithms could be benchmarked against this same limiting configuration
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

0 major / 3 minor

Summary. The paper proves asymptotic optimality of the first-fit discipline for dynamic storage allocation on the half-axis. Items arrive as a Poisson process of rate r with i.i.d. sizes from H (supported on a countable increasing sequence of positive reals, finite or infinite, with finite mean) and depart after unit-mean exponential times. The main result states that the r-scaled steady-state configuration under first-fit converges in distribution to the deterministic gap-free packing with items ordered by increasing size from left to right; as a corollary the scaled empty space vanishes.

Significance. The result resolves the 1985 conjecture of Coffman, Kadota and Shepp for non-degenerate H, extending their unit-size case. If the technical arguments hold, the work supplies the first rigorous proof of convergence in distribution to the ordered gap-free limit under a natural topology, thereby establishing asymptotic optimality via continuity of the empty-space functional. The countable-support assumption is explicitly scoped and the derivation is parameter-free.

minor comments (3)
  1. The abstract states the convergence claim and the required conditions on H, but the full derivation steps, technical lemmas, and handling of the continuous-time Markov process are not available for inspection in the provided excerpt; the central claim therefore cannot be verified from given text.
  2. Clarify the precise topology on the space of configurations under which convergence in distribution is proved (e.g., vague convergence on the half-axis or a suitable Skorokhod space) and confirm that the empty-space functional is continuous at the limit point.
  3. The statement that H is 'concentrated on countably many positive real sizes forming an increasing sequence' should be accompanied by an explicit definition of the support sequence (a_n) and the finite-mean condition in the main theorem statement.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive summary, significance assessment, and recommendation of minor revision. No major comments appear in the report, so we have no points requiring point-by-point response or manuscript changes.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; derivation self-contained

full rationale

The paper establishes convergence in distribution of the r-scaled steady-state first-fit configuration to an externally described deterministic limit (smaller items left, larger right, gap-free) for H supported on countable increasing sizes with finite mean. This limit is not defined via the result itself, nor obtained by fitting parameters or self-citation chains. The asymptotic optimality claim (vanishing scaled empty space) follows from continuity of the empty-space functional at the limit point. No load-bearing step reduces to a self-definition, fitted input renamed as prediction, or ansatz smuggled via prior self-work. The derivation is independent of the target claim and externally falsifiable under the stated assumptions.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The result rests on standard Poisson-arrival and exponential-service assumptions plus the countable-support restriction on H; no free parameters or new entities are introduced.

assumptions (3)
  • domain assumption Items arrive according to a Poisson process of rate r
    Standard modeling choice for dynamic storage allocation stated in the abstract.
  • domain assumption Service times are i.i.d. exponential with unit mean
    Common assumption for holding times, stated in the abstract.
  • domain assumption Item sizes are i.i.d. from H supported on a countable increasing sequence of positive reals with finite mean
    The key structural assumption on H that enables the proof for the non-degenerate case, stated in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Asymptotic optimality of dynamic first-fit packing on the half-axis." pith.science (2026). https://pith.science/paper/NPSF6WYE

@misc{pith2026240403797,
  author       = {Pith},
  title        = {Pith review of: Asymptotic optimality of dynamic first-fit packing on the half-axis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NPSF6WYE}},
  note         = {Machine review of arXiv:2404.03797}
}
abstract

We revisit a classical problem in dynamic storage allocation. Items arrive in a linear storage medium, modeled as a half-axis, at a Poisson rate $r$ and depart after an independent exponentially distributed unit mean service time. The arriving item sizes (lengths) are assumed to be independent and identically distributed (i.i.d.) from a common distribution $H$. A widely employed algorithm for allocating the items is the "first-fit" discipline, namely, each arriving item is placed in the left-most vacant interval large enough to accommodate it. In a seminal 1985 paper, Coffman, Kadota, and Shepp ([6]) proved that in the special case of unit length items (i.e. degenerate $H$), as $r$ tends towards infinity, the first-fit algorithm is asymptotically optimal in the following sense: the steady-state ratio of expected "empty space" (gaps between items) to expected occupied space tends towards $0$. In a sequel to [6], Coffman, Kadota, and Shepp ([5]) conjectured that the first-fit discipline is also asymptotically optimal for non-degenerate $H$. In this paper we provide the first proof of first-fit asymptotic optimality for non-degenerate distributions $H$ of item sizes. Our main result is for the case when $H$ is concentrated on countably many positive real sizes forming an increasing sequence that is either finite or goes to infinity, with the average item size being finite. We prove that under the first-fit discipline, as $r$ tends towards infinity, the steady-state packing configuration (scaled down by $r$) converges in distribution to the limiting packing configuration with smaller items on the left, larger items on the right, and with no gaps between. In particular, this proves asymptotic optimality of first-fit in the sense that in steady-state the empty space (scaled down by $r$) vanishes.

Figures

Figures reproduced from arXiv: 2404.03797 by the authors.

Figure 1
Figure 1. A simulation of the system state evolution for [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [1]

    Aldous, D. (1986). Some interesting processes arising as heavy traffic limits in an M/M/ ∞ storage process. Stochastic Processes and their Applications , 22(2), 291–313

  2. [2]

    Coffman, E.G., Flatto, L., and Leighton, F.T. (1990). First-fit allocation of queues: tight probabilistic bounds on wasted space. Stochastic Processes and their Applications , 36(2), 311–330

  3. [3]

    Coffman, E.G., Flatto, L., and Leighton, F.T. (1990). First-fit storage of linear lists: tight probabilis- tic bounds on wasted space. In Proceedings of the First Annual ACM-SIAM Symposium on Discrete Algorithms (pp. 272–279)

  4. [4]

    Coffman, E.G., Kadota, T.T., Leighton, F.T., and Shepp, L.A. (1986). Stochastic analysis of storage fragmentation. In Teletraffic Analysis and Computer Performance Evaluation (pp. 275–295)

  5. [5]

    Coffman, E.G., Kadota, T.T., and Shepp, L.A. (1985). On the asymptotic optimality of first-fit storage allocation. IEEE Transactions on Software Engineering , 11(2), 235–239

  6. [6]

    Coffman, E.G., Kadota, T.T., and Shepp, L.A. (1985). A stochastic model of fragmentation in dynamic storage allocation. SIAM Journal of Computing , 14(2), 416–425

  7. [7]

    and Leighton, F.T

    Coffman, E.G. and Leighton, F.T. (1986). A provably efficient algorithm for dynamic storage allocation. In Proceedings of the Eighteenth Annual ACM Symposium on Theory of Computing (pp. 77–90)

  8. [8]

    Knessl, C. (2000). Asymptotic expansions for a stochastic model of queue storage. The Annals of Applied Probability, 10(2), 592–615

Show all 14 references
  1. [9]

    Knessl, C. (2004). Some asymptotic results for the M/M/ ∞ queue with ranked servers. Queueing Sys- tems, 47, 201–250

  2. [10]

    and Lee, M.K

    Park, B.C. and Lee, M.K. (2007). Closest open location rule under stochastic demand. International Journal of Production Research, 45(7), 1695–1705

  3. [11]

    and Knessl, C

    Sohn, E. and Knessl, C. (2010). Storage allocation under processor sharing I: exact solutions and asymptotics. Queueing Systems, 65(1), 1–18

  4. [12]

    and Knessl, C

    Sohn, E. and Knessl, C. (2011). Storage allocation under processor sharing II: further asymptotic results. European Journal of Applied Mathematics , 22(1), 45–82

  5. [13]

    and Knessl, C

    Sohn, E. and Knessl, C. (2011). On some simple single server models of dynamic storage. In Proceedings of the 6th International Conference on Queueing Theory and Network Applications (pp. 161–168)

  6. [14]

    and Knessl, C

    Sohn, E. and Knessl, C. (2012). Some exact and asymptotic solutions to single server models of dynamic storage. Stochastic Models, 28(2), 248–280. 11

Pith tools

Reviewed May 24, 2026 · model on record in the stance chip above.