Pith. sign in

REVIEW 3 major objections 5 minor 27 references

Optimal Task Offloading with Firm Deadlines for Mobile Edge Computing Systems

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper proves that optimal task offloading in an edge system with firm deadlines is the smallest offloading step that reaches a non-offloading state, and that the infinite-state DP can be solved on a finite lean-state space with a…

desk verdict Genuinely useful structural results for a firm-deadline offloading DP, but the finite lean-state claim fails for T<N; an easy fix and a few small corrections should make this solid. read the letter →

arxiv 2506.09180 v1 pith:PNHEA7H4 submitted 2025-06-10 eess.SY cs.SY

Khai Doan , Wesley Araujo , Evangelos Kranakis , Ioannis Lambadaris , Yannis Viniotis , Wonjae Shin This is my paper
classification eess.SYcs.SY MSC 90C3990B36
keywords taskoffloadingfirmdeadlinesmobileedgecomputingdynamicprogrammingMarkovdecisionprocessleanstatespaceexcessivetaskspolicy
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

The paper studies a base station serving tasks with fixed deadlines by local processing (at most one task per slot, available with probability $\mu$) or by offloading to a remote server through an intermittently available autonomous mobile agent (available with probability $p_a$). Offloading costs $c_o$ per task, expiration costs $c_p > c_o$. The paper claims that the optimal policy has a crisp threshold form: offload the smallest number of most imminent tasks that makes the remaining state non-offloading, where a state is non-offloading exactly when removing the next most imminent task would change the optimal cost by less than $c_o$. The main structural result is that the infinite-state dynamic program collapses: every state's value equals its finite lean-state value plus a closed-form linear term (Proposition 1), so the DP needs to be solved only on finitely many lean states. If correct, this turns a seemingly intractable stochastic control problem into a finite computation plus an algebraic adjustment, with memory savings on the order of 95 to 98 percent in the paper's horizon-15 simulations.

What carries the argument

The load-bearing objects are the reduced states and lean states. A reduced state is one with no excessive tasks, the tasks guaranteed to expire before local processing can reach them, and the number of reduced states is the Catalan number. A lean state $s^{(\ell)}$ for a state $s$ is the finite representative that, under any AMA-free sequence of arrivals and local-service outcomes, reaches the same reduced state as $s$; the surplus tasks in $s$ are excessive and therefore act independently of the remaining tasks. The identity $V_T(s)=V_T(s^{(\ell)})+C_\ell$ (Proposition 1) carries the computational reduction, while discrete convexity of $F(s,F)=V^A_T(\bar{s}_F)+F c_o$ and the adjacent-state relation convert the policy into a threshold: offload until the residual state is non-offloading, where non-offloading is checked by the difference condition of Proposition 2.

What would settle it

Take a small instance, for example $N=3$, $T\in\{3,4,5\}$, $\mu=0.5$, $p_a=0.5$, $c_o=1$, $c_p=3$, and compute $V_T$ by exact DP on a sufficiently large truncated state space. Choose a state with surplus tasks, such as $s=(0,3,1)$ with lean state $s^{(\ell)}=(0,1,1)$, and compare $V_T(s)-V_T(s^{(\ell)})$ with the closed-form $C_\ell$ from Eq. (22); any mismatch would show that surplus tasks are not independent of the lean-state dynamics and that Proposition 1 fails.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that any state $s$ can be reduced to a finite lean state $s^{(\ell)}$ by deleting tasks that are guaranteed to expire unless offloaded (excessive tasks), and that $V_T(s)=V_T(s^{(\ell)})+C_\ell$, with $C_\ell$ a linear function of the surplus tasks (Eqs. 21–22). It then characterizes optimal decisions: for every state, the optimal number of tasks to offload is the smallest $F$ such that the state obtained by offloading the $F$ most imminent tasks is non-offloading (Theorem 2), and a state is non-offloading exactly when $V_T(s^{(a)})-V_T(s)<c_o$ for an adjacent state $s^{(a)}$ (Proposition 2). Together with discrete convexity of the cost in the offloading count (Lemma 2) and the fact that adjacent states' optimal decisions differ by at most one (Theorem 1), this yields a finite explicit policy: strip off guaranteed-to-expire tasks, solve a finite DP on the lean and reduced states, and read off the decision as the first step that reaches a non-offloading state.

Load-bearing premise

The reduction treats the surplus tasks that separate a state from its lean counterpart as fully non-interacting: each surplus task is either offloaded at the first AMA arrival or expires, and its presence never changes which non-surplus tasks are locally processed or offloaded.

Editorial extensions

If this is right

  • For any state, the optimal decision is to offload the smallest number of most imminent tasks that leaves a non-offloading state, which makes Algorithm II a simple iterative stop rule.
  • The DP equation needs to be evaluated only on the finite set of lean states; the value of every other state follows from Eqs. (21)–(22) by one linear algebra step.
  • Optimal decisions along chains of adjacent states can be inferred from a single known decision, because adjacent offloading counts differ by at most one and both are capped at zero.
  • In the paper's simulations, the lean-state transformation cuts the number of stored DP entries by roughly 95, 97, and 98 percent for state dimensions 3, 4, and 5 at horizon 15, with savings growing in dimension.
  • The structural results survive the addition of a local processing cost: only the DP equation changes, while the reduced-state, lean-state, and decision-rule framework remains valid.

Reading between the lines

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

  • If Proposition 1 extends to bulk arrivals as the paper's Remark 1 anticipates, the lean-state technique would likely apply to batch arrival processes, but the surplus-task independence would need to be verified because multiple simultaneous arrivals can change which tasks are most imminent.
  • The threshold rule suggests a direct online implementation: with the lean-state value table precomputed, a controller could make each slot's decision by comparing stored values rather than solving DP online, making the memory-computation trade-off explicit.
  • The condition $V_T(s^{(a)})-V_T(s)<c_o$ behaves like a shadow price for one extra task; a testable extension is to compute threshold curves in $p_a$, $\mu$, and the deadline distribution, which would give designers closed-form load limits.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper studies a discrete-time mobile edge computing system in which tasks with firm deadlines arrive at a base station and can either be processed locally (at most one task per slot, with probability mu) or offloaded to a remote server whenever an Autonomous Mobile Agent (AMA) is present (probability p_a). The authors formulate a finite-horizon dynamic program for minimizing the expected time-average cost, composed of offloading costs and task-expiration penalties, and derive structural properties of the optimal policy: discrete convexity of the cost as a function of the number of offloaded tasks, a relation between optimal decisions at adjacent states, and a characterization of the optimal decision as the smallest number of offloaded tasks that reaches a non-offloading state. The main computational contribution is Proposition 1, which claims that the minimum cost of any state can be obtained from a finite set of "lean states" via a linear correction term (Eq. 22), and Algorithm II uses this reduction to compute the optimal policy. The paper also reports numerical experiments on optimal decision visualization, memory savings, and comparisons with threshold, expiry-driven, random, and on-the-spot baselines.

Significance. If the finite lean-state reduction were established for all T and N, the paper would make a solid, self-contained contribution: it would convert an infinite-state DP into a finite computation plus an algebraic correction, give a structured optimal policy, and provide appendices with proofs of the main theorems. The structural insights, such as adjacent-state inference and the threshold characterization of offloading, are nontrivial and could inform richer MEC models. The paper's self-contained nature is a strength, and the simulations are consistent with the theory in the regime T >= N used in the figures. However, the central finiteness claim is currently false as stated for T < N (see major comment 1), and the proof of the key separation property behind Proposition 1 is asserted rather than demonstrated (see major comment 2). Because these points are load-bearing, the paper is not yet acceptable in its present form, though the issues appear fixable by an explicit assumption or a truncated-state re-derivation.

major comments (3)
  1. [Section IV-B, Definition 2 and Algorithm I] The claimed finiteness of the lean state space fails when T < N. In Definition 2, gamma_i = 0 for every i > min(N,T), and Algorithm I loops only up to min(N,T) (line 4). Consequently, for such i, b_i^r = b_i and Eq. (20) gives b_i^ell = max(0, b_i^r) = b_i. Hence every state of the form (0, ..., 0, b_{T+1}, ..., b_N) is its own lean state; for example, with N = 2 and T = 1, the states (0,k) for k = 0,1,2,... are all lean, so the set is infinite. This contradicts the statement after Eq. (21) that the lean state space is finite, the memory argument for Algorithm II, and the memory-saving claims in Section VII-C, whose Fig. 7 only covers T >= N. The paper should either add an explicit T >= N assumption or truncate the state representation to min(N,T) components (tasks with deadline greater than T cannot affect the cost in Eq. (11)) and re-derive the lean-state construction and Eq. (21) on that truncated space.
  2. [Appendix B, Eqs. (41)-(48)] The proof of Proposition 1 assumes the key separation property it needs to prove. It asserts that the extra tasks d_i = b_i - b_i^ell are "excessive tasks" and "should be offloaded by the optimal policy" without showing from the transition law that these surplus tasks never change which non-excessive tasks are locally processed or offloaded. The case analysis in Eqs. (41)-(48) postulates that, under every realization, the remaining state of s after serving or expiring the surplus tasks is exactly the corresponding state of s^ell, plus the same arrivals and local service. If surplus tasks can occupy the local server or the offloading budget at a time when s^ell would serve a different task, Eq. (21) would not follow. A formal coupling or monotonicity argument is needed; as written, this is a gap in the central reduction.
  3. [Section V-A, Eq. (24)] The domain of F(s,F) in Eq. (24) is inconsistent with the text immediately above it and with the appendix's L1(s) in Eq. (33). The text says that decisions F < b1 are removed because deadline-1 tasks are excessive; for b1 = 0 this set is empty. Yet Eq. (24) removes {0} through the term max(b1 - 1, 0). Consequently, Lemma 2's stated domain excludes the candidate F = 0 for states with b1 = 0, and Lemma 3 cannot hold for non-offloading states with b1 = 0. The correct domain appears to be L(s) \ {0, ..., b1 - 1}, matching L1(s). The appendix proofs use L1(s), but the formal statement of Lemma 2 must be corrected.
minor comments (5)
  1. [Algorithm II, line 20] In the pseudocode, the stored quadruplet is written with horizon T rather than T - t; the surrounding text and Fig. 7 use the latter, so the pseudocode should be aligned with the description.
  2. [Appendix E] The phrase "convexity of cost functions proven in Subsec. VIII" should refer to the relevant appendix section (Appendix C) or the main-text Section V-A, not "Subsec. VIII."
  3. [Algorithm II, line 15] Line 15 says "Offloading F_r most imminent tasks from s(r)"; as written this does not define s* correctly. It should state that s* is obtained by offloading F* = F_r + F_g most imminent tasks from the original state s.
  4. [Remark 1] The claim that Eq. (21) extends to bulk task arrivals is stated without proof. If kept, the authors should either provide a proof or state precisely how the arrival process is modified and why the separation property remains valid.
  5. [Section VII-C] The memory-saving experiments in Fig. 7 use T = 6, ..., 15 and N = 3, 4, 5, i.e., only the regime T >= N. The paper should state this limitation explicitly, or extend the experiments to T < N if the truncated-state fix is adopted.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; minor self-citation to conference version is not load-bearing.

full rationale

Proposition 1's linear relation (Eq. 21) is not true by construction: the lean state in Definition 2 is defined through gamma_i and the reduced state, not through V_T(s) - V_T(s^ell). Appendix B derives C_ell from a case analysis over the first AMA arrival time; whether surplus tasks are truly independent is a substantive proof assumption, not an input-output equivalence. Theorem 2 is likewise derived from Lemma 2 and Theorem 1 in Appendices C-G rather than being a restatement of Definition 5. The only self-citation, [24], is a conference version of the same model, and the present paper contains the proofs, so no load-bearing claim is imported from it. The simulations compare against external baselines [26], [27] with fixed parameters. One correctness caveat, the T<N case leaves gamma_i=0 for i>min(N,T) so trailing components make the lean set infinite, is a gap in the paper's finiteness claim, but it is not circularity: Eq. (21) is not equivalent to its own input in that regime; it is a potentially false structural assertion. No specific circular step is therefore identified.

Assumptions & free parameters 0 free parameters · 6 assumptions · 1 invented entities

No parameters are fitted to data for the central derivation; all numeric inputs to the DP are exogenous model parameters. The main unproved premises are the Bernoulli and independence assumptions for AMA presence, arrivals, and local service, the single-arrival and single-service-per-slot capacity, the ascending-deadline service order, and the additivity of surplus tasks in the lean-state mapping. The AMA is a modeling abstraction, not an independently evidenced entity.

assumptions (6)
  • domain assumption AMA presence, task arrivals, and local processing opportunities are independent Bernoulli events in each time slot.
    Section II.A.1, II.A.3, and II.A.4 define these probabilities and independence; all DP transition probabilities in Eq. (13) rely on them.
  • domain assumption At most one task arrives per slot and at most one task is locally processed per slot.
    Section II.A.3-4; this bound is used to define excessive tasks and to obtain the Catalan count of reduced states in Lemma 1.
  • domain assumption Offloading and local processing always target the most imminent tasks first.
    Section II.B states this is trivially optimal because all tasks share identical costs; the structural results Theorem 1-2 and Lemma 2 assume this fixed service order.
  • domain assumption The offloading cost is below the expiration penalty, co < cp.
    Section II.A.2; otherwise offloading excessive tasks is never beneficial and the reduced-state and lean-state construction changes.
  • domain assumption Tasks with deadlines exceeding the remaining horizon are never offloaded and do not affect the optimal cost.
    Section IV.A uses this to stop Algorithm I at min(D,T); it follows from there being no terminal cost and co > 0, but it is not proven in the paper.
  • domain assumption Surplus tasks removed in the lean-state mapping are non-interfering: they are offloaded at first AMA arrival or expire without altering the evolution of the lean-state tasks.
    Appendix B, Eqs. (41)-(48) decompose the cost case by case under this separation; it is the load-bearing premise of Proposition 1.
invented entities (1)
  • Autonomous Mobile Agent (AMA) relay
    purpose: Provides intermittent connectivity between the base station and the remote server; offloading is only possible in slots when the AMA is present with probability a.
    Standard modeling abstraction for random availability; the paper attaches no falsifiable prediction to the AMA outside the model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Task Offloading with Firm Deadlines for Mobile Edge Computing Systems." pith.science (2026). https://pith.science/paper/PNHEA7H4

@misc{pith2026250609180,
  author       = {Pith},
  title        = {Pith review of: Optimal Task Offloading with Firm Deadlines for Mobile Edge Computing Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PNHEA7H4}},
  note         = {Machine review of arXiv:2506.09180}
}
read the original abstract

Under a dramatic increase in mobile data traffic, a promising solution for edge computing systems to maintain their local service is the task migration that may be implemented by means of Autonomous mobile agents (AMA). In designing an optimal scheme for task offloading to AMA, we define a system cost as a minimization objective function that comprises two parts. First, an offloading cost which can be interpreted as the cost of using computational resources from the AMA. Second, a penalty cost due to potential task expiration. To minimize the expected (timeaverage) cost over a given time horizon, we formulate a Dynamic programming (DP). However, the DP Equation suffers from the well-known curse of dimensionality, which makes computations intractable, especially for infinite system state space. To reduce the computational burden, we identify three important properties of the optimal policy and show that it suffices to evaluate the DP Equation on a finite subset of the state space only. We then prove that the optimal task offloading decision at a state can be inferred from that at its adjacent states, further reducing the computational load. We present simulations to verify the theoretical results and to provide insights into the considered system.

Figures

Figures reproduced from arXiv: 2506.09180 by the authors.

Figure 1
Figure 1. System model illustration where user tasks arrive at [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Illustration of events occurring in a time slot. Amon [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Illustrative example of the concept of lean states. I [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Visualisation of Theorem 2 for the state vector [PITH_FULL_IMAGE:figures/full_fig_p027_4.png]
Figure 5
Figure 5. Figure 5: Visualisation of Theorem 2 for the state vector [PITH_FULL_IMAGE:figures/full_fig_p027_5.png]
Figure 6
Figure 6. Figure 6: Visual interpretation of Eqs. (30)-(31) and Theorem [PITH_FULL_IMAGE:figures/full_fig_p030_6.png]
Figure 7
Figure 7. Figure 7: Memory savings achieved by using the transformation [PITH_FULL_IMAGE:figures/full_fig_p032_7.png]
Figure 8
Figure 8. Figure 8: Overall cost (defined in Eq. (6)) versus the local processing probability, , for the proposed optimal algorithm, threshold method [26], ED, and random methods. For each value of , the optimal threshold is used for the threshold method. The system parameter configuratio…
Figure 9
Figure 9. Figure 9: Evaluation of the average cost per task incurred and the percentage of available local processing services that can be used by the proposed method, the threshold method [26] at its optimal threshold value, and the OTS scheme [27]. We note that the ED and Random methods…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

  1. [24]

    Optimal Task Offloading Policy in Edge Computing Systems with Firm Deadlines,

    K. Doan, W. Araujo, E. Kranakis, I. Lambadaris, and Y. Vi niotis, “Optimal Task Offloading Policy in Edge Computing Systems with Firm Deadlines,” in Proc. IEEE Global Telecommun. Conf. , Dec. 2023, pp. 916–921

    work page 2023

  2. [1]

    A Survey on Mobile Edge Computing: The Communication Pers pective,

    Y. Mao et al. , “A Survey on Mobile Edge Computing: The Communication Pers pective,” IEEE Commun. Surv. Tutor., vol. 19, no. 4, pp. 2322–2358, 2017

    work page 2017

  3. [2]

    The Extended Cloud: Review and Analysis of Mobile Edge Computing and Fog From a Security and Resilience Perspe ctive,

    S. N. Shirazi, A. Gouglidis, A. Farshad, and D. Hutchison , “The Extended Cloud: Review and Analysis of Mobile Edge Computing and Fog From a Security and Resilience Perspe ctive,” IEEE J. Sel. Areas Commun. , vol. 35, no. 11, pp. 2586–2595, 2017

    work page 2017

  4. [3]

    Task Offloading and Resource Allocation for Mobile Edge Computing by Deep Reinforcement Learning Ba sed on SARSA,

    T. Alfakih, M. M. Hassan, A. Gumaei, C. Savaglio, and G. Fo rtino, “Task Offloading and Resource Allocation for Mobile Edge Computing by Deep Reinforcement Learning Ba sed on SARSA,” IEEE Access, vol. 8, pp. 54 074–54 084, 2020. 53

    work page 2020

  5. [4]

    A Survey of Mobi le Cloud Computing: Architecture, Applications, and Approaches,

    H. T. Dinh, C. Lee, D. Niyato, and P. Wang, “A Survey of Mobi le Cloud Computing: Architecture, Applications, and Approaches,” Wirel. Commun. Mob. Comput. , vol. 13, no. 18, pp. 1587–1611, 2013

    work page 2013

  6. [5]

    A Sur vey of Mobile Cloud Computing Application Models,

    A. u. R. Khan, M. Othman, S. A. Madani, and S. U. Khan, “A Sur vey of Mobile Cloud Computing Application Models,” IEEE Commun. Surv. Tutor. , vol. 16, no. 1, pp. 393–413, 2014

    work page 2014

  7. [6]

    Dynamic Operations of C loud Radio Access Networks (C-RAN) for Mobile Cloud Computing Systems,

    M. Y.-K. Chua, F. R. Yu, and S. Bu, “Dynamic Operations of C loud Radio Access Networks (C-RAN) for Mobile Cloud Computing Systems,” IEEE Trans. Veh. Technol. , vol. 65, no. 3, pp. 1536–1548, 2016

    work page 2016

  8. [7]

    DATS: Disper sive Stable Task Scheduling in Heterogeneous Fog Networks,

    Z. Liu, X. Yang, Y. Yang, K. Wang, and G. Mao, “DATS: Disper sive Stable Task Scheduling in Heterogeneous Fog Networks,” IEEE Internet Things J. , vol. 6, no. 2, pp. 3423–3436, 2019

    work page 2019

Show all 27 references
  1. [8]

    Fog Computing Ar chitecture, Evaluation, and Future Research Directions,

    M. Aazam, S. Zeadally, and K. A. Harras, “Fog Computing Ar chitecture, Evaluation, and Future Research Directions,” IEEE Commun. Mag. , vol. 56, no. 5, pp. 46–52, 2018

    2018

  2. [9]

    Survey of Fog Computin g: Fundamental, Network Applications, and Research Challenges,

    M. Mukherjee, L. Shu, and D. Wang, “Survey of Fog Computin g: Fundamental, Network Applications, and Research Challenges,” IEEE Commun. Surv. Tut. , vol. 20, no. 3, pp. 1826–1857, 2018

    2018

  3. [10]

    QoE and Power Efficiency Tradeoff for Fog Computing Networks with Fog Node Cooperation,

    Y. Xiao and M. Krunz, “QoE and Power Efficiency Tradeoff for Fog Computing Networks with Fog Node Cooperation,” in Proc. IEEE Conf. Comput. Commun. , 2017, pp. 1–9

    2017

  4. [11]

    A Su rvey on Computation Offloading Modeling for Edge Computing,

    H. Lin, S. Zeadally, Z. Chen, H. Labiod, and L. Wang, “A Su rvey on Computation Offloading Modeling for Edge Computing,” J. Netw. Comput. Appl. , vol. 169, p. 102781, 2020

    2020

  5. [12]

    Edge Cloud Offloading Algorithms: Issues, Methods, and Perspectives,

    J. Wang, J. Pan, F. Esposito, P. Calyam, Z. Yang, and P. Mo hapatra, “Edge Cloud Offloading Algorithms: Issues, Methods, and Perspectives,” ACM Comput. Surv. , vol. 52, no. 1, Feb 2019

    2019

  6. [13]

    Collaborative Task Execu tion in Mobile Cloud Computing Under a Stochastic Wireless Channel,

    W. Zhang, Y. Wen, and D. O. Wu, “Collaborative Task Execu tion in Mobile Cloud Computing Under a Stochastic Wireless Channel,” IEEE Trans. Wirel. Commun. , vol. 14, no. 1, pp. 81–93, 2014

    2014

  7. [14]

    Deadline-Aware Task Offload ing with Partially-Observable Deep Reinforcement Learning for Multi-Access Edge Computing,

    H. Huang, Q. Ye, and Y. Zhou, “Deadline-Aware Task Offload ing with Partially-Observable Deep Reinforcement Learning for Multi-Access Edge Computing,” IEEE Trans. Netw. Sci. Eng. , 2021

    2021

  8. [15]

    An Optimization-Based Approa ch to Offloading in Ad-Hoc Mobile Clouds,

    D. Van Le and C.-K. Tham, “An Optimization-Based Approa ch to Offloading in Ad-Hoc Mobile Clouds,” in Proc. IEEE Global Telecommun. Conf. , Dec. 2017, pp. 1–6

    2017

  9. [16]

    Computation Offloadin g and Resource Allocation in Mixed Fog/Cloud Computing Systems with Min-Max Fairness Guarantee,

    J. Du, L. Zhao, J. Feng, and X. Chu, “Computation Offloadin g and Resource Allocation in Mixed Fog/Cloud Computing Systems with Min-Max Fairness Guarantee,” IEEE Trans. Commun. , vol. 66, no. 4, pp. 1594–1608, 2017

    2017

  10. [17]

    A Deep Reinforcement Learning Based Offloading Scheme in Ad-Hoc Mobile Clouds,

    D. Van Le and C.-K. Tham, “A Deep Reinforcement Learning Based Offloading Scheme in Ad-Hoc Mobile Clouds,” in Proc. IEEE Conf. Comput. Commun. Wkshps. , Apr. 2018, pp. 760–765

    2018

  11. [18]

    A rtificial Intelligence Enabled Distributed Edge Computing for Internet of Things Applications,

    G. Fragkos, E. E. Tsiropoulou, and S. Papavassiliou, “A rtificial Intelligence Enabled Distributed Edge Computing for Internet of Things Applications,” in Proc. IEEE Distrib. Comput. Sensor Syst. , May 2020, pp. 450–457

    2020

  12. [19]

    Multiobjective Optim ization for Joint Task Offloading, Power Assignment, and Resource Allocation in Mobile Edge Computing,

    P. Wang, K. Li, B. Xiao, and K. Li, “Multiobjective Optim ization for Joint Task Offloading, Power Assignment, and Resource Allocation in Mobile Edge Computing,” IEEE Internet Things J. , vol. 9, no. 14, pp. 11 737–11 748, 2022

    2022

  13. [20]

    Computati on Offloading in Multi-Access Edge Computing: A Multi-Task Learning Approach,

    B. Yang, X. Cao, J. Bassey, X. Li, and L. Qian, “Computati on Offloading in Multi-Access Edge Computing: A Multi-Task Learning Approach,” IEEE Trans. Mobile Comput. , vol. 20, no. 9, pp. 2745–2762, 2021

    2021

  14. [21]

    Delay Constrained Hybrid Ta sk Offloading of Internet of Vehicle: A Deep Reinforcement Learning Method,

    C. Wu, Z. Huang, and Y. Zou, “Delay Constrained Hybrid Ta sk Offloading of Internet of Vehicle: A Deep Reinforcement Learning Method,” IEEE Access, vol. 10, pp. 102 778–102 788, 2022

    2022

  15. [22]

    A Multiagent Met a-Based Task Offloading Strategy for Mobile-Edge Computing,

    W. Ding, F. Luo, C. Gu, Z. Dai, and H. Lu, “A Multiagent Met a-Based Task Offloading Strategy for Mobile-Edge Computing,” IEEE Trans. Cogn. Develop. Sys. , vol. 16, no. 1, pp. 100–114, 2024

    2024

  16. [23]

    Inte lligent Task Offloading and Resource Allocation in 54 Digital Twin Based Aerial Computing Networks,

    H. Guo, X. Zhou, J. Wang, J. Liu, and A. Benslimane, “Inte lligent Task Offloading and Resource Allocation in 54 Digital Twin Based Aerial Computing Networks,” IEEE J. Sel. Areas Commun. , vol. 41, no. 10, pp. 3095–3110, 2023

    2023

  17. [25]

    R. P. Stanley, Catalan numbers. Cambridge University Press, 2015

    2015

  18. [26]

    Distributed Threshold-base d Offloading for Large-Scale Mobile Cloud Computing,

    X. Qin, B. Li, and L. Ying, “Distributed Threshold-base d Offloading for Large-Scale Mobile Cloud Computing,” in Proc. IEEE Conf. Comput. Commun. , 2021, pp. 1–10

    2021

  19. [27]

    Mobile Data O ffloading: How Much Can WiFi Deliver?

    K. Lee, J. Lee, Y. Yi, I. Rhee, and S. Chong, “Mobile Data O ffloading: How Much Can WiFi Deliver?” IEEE/ACM Trans. Netw. , vol. 21, no. 2, pp. 536–550, 2013. Khai Doan received his PhD in Information Systems Technology and Desi gn from Singapore University of Technology and Des...

    2000

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.