Pith. sign in

REVIEW 4 major objections 4 minor 29 references

A Distributed Gradient-Based Deployment Strategy for a Network of Sensors with a Probabilistic Sensing Model

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read A distributed gradient rule lets mobile sensors raise coverage using only local information.

desk verdict A plausible engineering extension with a real gap in the gradient derivation and unproven convergence claims; the simulations show useful coverage gains, but the theory needs tightening before publication. read the letter →

arxiv 2509.02869 v1 pith:BCBTXPT7 submitted 2025-09-02 eess.SY cs.SY

classification eess.SYcs.SY
keywords mobilesensordeploymentprobabilisticsensingElfesmodelVoronoipartitiongradient-basedoptimizationcoveragecontrolobstaclevisibilityhybridwirelessnetwork
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 addresses a practical question in wireless sensor networks: when sensors have probabilistic, distance-decaying detection and obstacles block sensing, how can mobile sensors reposition to improve overall coverage without a central coordinator? The authors claim that by partitioning the region into local Voronoi cells, the global coverage objective can be split into local terms, and each mobile sensor can ascend the gradient of its own local term using only information from nearby sensors. Their rule moves a sensor only when the predicted local gain exceeds a threshold, uses a step size that first grows then decays, and projects illegal positions back to the visible feasible region. If correct, this gives a scalable, energy-conscious, obstacle-aware deployment method that improves on static placement and converges to a configuration where no sensor and its neighbors can improve by moving.

What carries the argument

The load-bearing object is the Voronoi-based decomposition of coverage into local contributions, specifically Lemma 2's identity (13) asserting that the mobile-sensor-covered regions are disjoint and jointly exhaust the area covered by mobile sensors, so that maximizing each local Fi also maximizes the global objective F. The gradient computation (15)-(16) converts the change in a local region into three computable integrals, and the step-size rule (18) plus the epsilon threshold convert that gradient into guaranteed, energy-conscious movement. The projection step keeps each sensor inside its assigned region and in line of sight.

What would settle it

Take a small network with two mobile sensors whose sensing ranges overlap but whose communication radius is smaller than the distance between them, so they are not neighbors, plus one stationary sensor. Compute the true global coverage F directly by integrating φ(q)·max_i psi(q), then compute the right-hand side of equation (13); any positive discrepancy shows the decomposition and therefore the gradient's target undercounts coverage. Alternatively, place one obstacle so that the closest sensor to a point q is occluded while a farther sensor has line of sight: at that point ps(q) > psi(q) for

Watch

Extended reading notes

Core claim

Using the Elfes probabilistic sensing model, the paper defines network coverage as an integral of the maximum detection probability, then uses Voronoi partitioning to rewrite it as the sum of local coverage terms. The central formula is the gradient of a mobile sensor's local coverage (Theorem 1, equations 15-16): a surface integral over the overlap of the sensor's sensing disk with its local region, a line integral over the moving arc of the sensing boundary, and additional line integrals for obstacle-vertex cut-offs. All of these rely only on neighbor positions and obstacle geometry. Algorithm 1 then performs distributed gradient ascent on this local term with a dynamic step size, projecti

Load-bearing premise

The load-bearing premise is that the coverage decomposition (13) is exact: the local mobile sensing regions are disjoint and jointly exhaust the area covered by mobile sensors, and the nearest-sensor identity in Lemma 1 holds even when obstacles block line of sight. If overlap or visibility breaks this decomposition, the gradient climbs a surrogate objective rather than the true coverage.

Editorial extensions

If this is right

  • If the decomposition and gradient ascent are sound, each mobile sensor can decide its motion using only information from within communication range, so the strategy scales to large networks without a central planner.
  • Because movement is accepted only when the local coverage gain exceeds epsilon, the algorithm naturally stops when marginal gains are small, saving energy and giving a concrete stopping rule.
  • The obstacle terms in the gradient and the projection step together let sensors route around obstacles while preserving existing communication links.
  • Starting from a connected configuration, the total coverage never decreases and the algorithm terminates in finite time, giving predictable deployment behavior.
  • In the paper's simulation with 30 mobile and 5 stationary sensors, the rule raises area coverage from about 27% to 90%, indicating the practical magnitude of improvement.

Reading between the lines

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

  • The authors leave implicit that the step-size rule (18) depends on the gradient norm in the exponent, so sensors in tiny local regions near obstacle vertices may take very different step sizes than sensors in open areas; a direct numerical comparison with the alternative rule they mention (ηt = min(ηmax, η0/(φ(xi)||∇Fi||))) would test whether the chosen rule is the better trade-off.
  • Lemma 2 assumes the mobile local regions are disjoint even when two mobile sensors whose sensing ranges overlap are not communication neighbors. A simple two-sensor counterexample with communication radius smaller than the overlap distance would show whether the right-hand side of (13) equals the true coverage F; if not, the local gradient is climbing a surrogate objective.
  • The visibility constraint appears in D(si) but is not used in Lemma 1's proof, so a point q that is nearest to an occluded sensor but visible to a farther one is a place where the assigned sensor probability differs from the true maximum; the authors could quantify this discrepancy in a one-obstacle simulation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. This paper proposes a distributed deployment algorithm for a hybrid wireless sensor network in which sensing follows the Elfes probabilistic model and obstacles block line of sight. The authors partition the ROI using Voronoi concepts, define a local coverage integral F_i over a locally assigned region, compute a gradient formula (15)-(16), and move each mobile sensor along that gradient with a dynamically scheduled step size, accepting a move only if F_i increases by at least epsilon. They claim the global coverage F can be decomposed into local terms, that local optimization of F_i maximizes F, and that Algorithm 1 converges to a steady-state local-maximum configuration. Two simulations show improved area and weighted coverage relative to the initial static deployment.

Significance. If the theoretical claims were sound, the paper would provide a practical, distributed algorithm for a realistic coverage problem with probabilistic sensing and obstacles. The two simulation examples are encouraging as a proof of concept, and the step-size schedule with an acceptance threshold is a sensible engineering choice. The paper also builds on the authors' prior work, citing the relevant sources. However, the correctness of the central mechanism is not established: the coverage decomposition and the gradient derivation contain gaps that affect the main optimality and convergence claims. The paper is not yet ready for publication in its current form.

major comments (4)
  1. [Section III, Lemma 1 and Eq. (10)] Lemma 1 is false in the presence of obstacles. The proof uses only Euclidean distance monotonicity and ignores the visibility constraint Phi(x_i) in Eq. (2). A point q can be closer to x_i than to all other sensors while x_i is occluded from q; then psi_i(q)=0, but a farther visible sensor may have psi_j(q)>0, so p_s(q) != psi_i(q). Consequently, Eq. (10) and the stationary term in Eq. (13) do not represent the true coverage F.
  2. [Section III, Lemma 2 and Eq. (13)] The assertion that the mobile local regions Pi'_i ∩ D(s_i) are disjoint and jointly exhaust the mobile contribution is not proved and is false without an explicit assumption linking r_c to the sensing radii. If r_c < 2 r_s^max, two sensors can have overlapping D(s_i) without being communication neighbors; then a point in the overlap can be assigned to both Pi'_i and Pi'_j. Similarly, a stationary sensor outside the mobile sensor's neighbor set can dominate a point, so the stationary and mobile terms can double-count. The proof's 'by construction' sentence is not a derivation. Add an explicit assumption (e.g., r_c >= 2 r_s^max) and prove disjointness, or revise the decomposition.
  3. [Section IV, Theorem 1, Eqs. (15)-(16)] The claimed gradient is incomplete. Applying the Leibniz rule to F_i = ∫_{Pi'_i ∩ D(s_i)} φ psi_i dq yields a boundary integral over ∂(Pi'_i ∩ D(s_i)), which includes the moving Voronoi edges ∂Pi'_i inside D(s_i). No such integral appears in (16a)-(16c); the boundary components listed before Theorem 1 explicitly mention edges of Pi_i, but the formula drops them. Unlike in the global coverage sum, these single-cell boundary terms do not cancel. Therefore (15) is not ∇_{x_i} F_i, and Algorithm 1 is not a gradient ascent on F_i. This breaks the local-maximum interpretation of the stopping condition.
  4. [Section IV, Remark 3] The claim that F is non-decreasing and converges in finite time depends on Lemma 2 and on the candidate direction being an ascent direction for the true local objective. Since Lemma 2 is unproved and the gradient is incomplete, a move that increases the surrogate F_i need not increase F. The finite-time convergence argument therefore lacks a valid Lyapunov function. Either restore an exact decomposition and an exact gradient, or reframe the algorithm as a heuristic and remove the convergence claims.
minor comments (4)
  1. [Throughout] 'V oronoi' should be 'Voronoi' (spacing error). Also, 'complexifies' in Section I is informal.
  2. [Section IV, Eq. (18)] The norm in the step-size formula is typeset incorrectly ('|∇xi Fi∥' should be '∥∇xi Fi∥'), and the expression η0 t e^{-βt} ∥∇xi Fi∥ looks dimensionally odd: it mixes the gradient norm with a time-dependent factor. Clarify the intended units and check the formula.
  3. [Section IV, before Theorem 1] The statement 'the boundary of this region is assumed to comprise...' is an assumption, not a derivation. Please justify why these are the only boundary segments, especially in view of the moving Voronoi edges.
  4. [Section V] The simulation parameters are chosen by trial and error, and no comparison with existing algorithms or statistical variation over multiple random initializations is provided. A sensitivity analysis or at least error bars over several runs would make the empirical claims more robust.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the derivation is based on the problem setup and peer-reviewed external gradient formulas, with no fitted input passed off as a prediction.

full rationale

The paper's central derivation is not circular. The coverage reformulation in (13) is argued in-text from the Voronoi partition and the definition of Pi'_i; the disjointness claim is asserted rather than proved, and Lemma 1 omits the visibility function Phi(x_i) from (2), but these are correctness/rigor gaps, not instances of a conclusion being identical to an input. The local gradient formula (15)-(16) is presented as a theorem; the only external import is the boundary-equivalence statement "It is shown in [10], [12] that these terms are equivalent to (16b) and (16c), respectively," and those are published, independently checkable papers rather than concealed ansatze or fitted parameters. No constant is fitted to a target coverage value and then renamed a prediction: the algorithm and threshold rule directly evaluate the same coverage objective F, so the simulations demonstrate optimization rather than reverse-engineer a result. The potential missing Voronoi-edge boundary term in (15) is a mathematical-correctness issue about whether Algorithm 1 is true gradient ascent; it does not make the derivation circular. Consequently, the self-citations are visible and non-load-bearing for the core decomposition, and no circular step satisfies the evidentiary standard.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The algorithm's central claim rests on the exactness of the Voronoi-based coverage decomposition, which requires several unproven or boundary assumptions; the four algorithm parameters (eta0, beta, etamax, epsilon) are explicitly tuned by trial and error; sensing model parameters (rmin, rmax, alpha) and network sizes are problem inputs rather than fitted values. No physical entities are invented; the dynamic step size and threshold rule are algorithmic constructs.

free parameters (4)
  • eta0 (base step size) = 0.1
    Tuned by trial and error (Section V); controls exploration magnitude, large values cause premature convergence.
  • beta (step-size decay rate) = 0.04
    Tuned by trial and error (Section V); sets the peak time t=1/beta of the step envelope.
  • etamax (maximum step) = 2
    Hand-chosen bound on per-iteration displacement; invoked in the connectivity argument of Remark 2.
  • epsilon (movement acceptance threshold) = 10^-3
    Hand-chosen; trades precision against number of iterations and is essential to the finite-time convergence claim (Remark 3).
assumptions (5)
  • domain assumption Nearest sensor has the highest sensing probability over its Voronoi cell even with visibility constraints (Lemma 1)
    Requires identical sensors and monotone decay of psi with distance, but the proof ignores Phi(xi) in (2); the statement is false when the nearest sensor is occluded and a farther sensor has line of sight (Section III, Lemma 1).
  • ad hoc to paper The local mobile regions Pi'_i and D(si) are disjoint and exhaust the mobile-covered area (Lemma 2)
    Asserted without proof; can fail when two mobile sensors with overlapping sensing ranges are not communication neighbors, because each builds its Voronoi cell from its own local view (Section III, Lemma 2, Eq. (13)).
  • domain assumption Obstacles block sensing but not communication; the communication graph is an undirected disk graph of radius rc
    Used to define Ni in (4) and to claim connectivity preservation in Remark 2 (Section II-B).
  • standard math Leibniz integral rule applies to the moving-boundary local coverage integral, with boundary terms exactly as in [10], [12]
    Invoked in the proof of Theorem 1 via [29]; the boundary decomposition (16b)-(16c) is delegated to the authors' earlier work rather than re-derived for the Elfes model (Section IV, Theorem 1).
  • ad hoc to paper Every accepted move increases the global objective F, so F is non-decreasing and converges in finite time (Remark 3)
    Asserted, not proven; the acceptance rule evaluates Fi under the current partition, but the partition changes with movement, so a local gain does not imply a global gain (Section IV, Algorithm 1, Remark 3).

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Distributed Gradient-Based Deployment Strategy for a Network of Sensors with a Probabilistic Sensing Model." pith.science (2026). https://pith.science/paper/BCBTXPT7

@misc{pith2026250902869,
  author       = {Pith},
  title        = {Pith review of: A Distributed Gradient-Based Deployment Strategy for a Network of Sensors with a Probabilistic Sensing Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BCBTXPT7}},
  note         = {Machine review of arXiv:2509.02869}
}
read the original abstract

This paper presents a distributed gradient-based deployment strategy to maximize coverage in hybrid wireless sensor networks (WSNs) with probabilistic sensing. Leveraging Voronoi partitioning, the overall coverage is reformulated as a sum of local contributions, enabling mobile sensors to optimize their positions using only local information. The strategy adopts the Elfes model to capture detection uncertainty and introduces a dynamic step size based on the gradient of the local coverage, ensuring movements adaptive to regional importance. Obstacle awareness is integrated via visibility constraints, projecting sensor positions to unobstructed paths. A threshold-based decision rule ensures movement occurs only for sufficiently large coverage gains, with convergence achieved when all sensors and their neighbors stop at a local maximum configuration. Simulations demonstrate improved coverage over static deployments, highlighting scalability and practicality for real-world applications.

Figures

Figures reproduced from arXiv: 2509.02869 by the authors.

Figure 1
Figure 1. Comparison of sensing models: (a) deterministic [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Sample sensor configuration in a local coverage [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Initial configuration of the WSN in Example 1 [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Final configuration of the WSN in Example 1 [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 6
Figure 6. Figure 6: Quality of sensing in the initial configuration of the [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Quality of sensing in the final configuration of the [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 27 canonical work pages

  1. [1]

    Wireless sensor networks: a survey,

    I. F. Akyildiz, W. Su, Y . Sankarasubramaniam, and E. Cayirci, “Wireless sensor networks: a survey,” Computer networks , vol. 38, no. 4, pp. 393–422, 2002

  2. [2]

    The comprehensive gateway model for diverse environmental monitoring upon wireless sensor network,

    H.-C. Lin, Y .-C. Kan, and Y .-M. Hong, “The comprehensive gateway model for diverse environmental monitoring upon wireless sensor network,” IEEE Sensors Journal , vol. 11, no. 5, pp. 1293–1303, 2011

  3. [3]

    A military surveillance system based on wireless sensor networks with extended coverage life,

    C. V . Mahamuni, “A military surveillance system based on wireless sensor networks with extended coverage life,” in Proceedings of Inter- national conference on global trends in signal processing, information computing and communication (ICGTSPICC) , pp. 375–381, 2016

  4. [4]

    A survey on clustering algorithms for wireless sensor networks,

    A. A. Abbasi and M. Younis, “A survey on clustering algorithms for wireless sensor networks,” Computer communications, vol. 30, no. 14- 15, pp. 2826–2841, 2007

  5. [5]

    Sensor deployment and target localization based on virtual forces,

    Y . Zou and K. Chakrabarty, “Sensor deployment and target localization based on virtual forces,” in Proceedings of IEEE INFOCOM. 22nd Annual Joint Conference of the IEEE Computer and Communications Societies, vol. 2, pp. 1293–1303, 2003

  6. [6]

    Obstacle detection and estimation in wire- less sensor networks,

    W.-T. Wang and K.-F. Ssu, “Obstacle detection and estimation in wire- less sensor networks,” Computer Networks, vol. 57, no. 4, pp. 858–868, 2013

  7. [7]

    Coverage protocols for wireless sensor networks: Review and future directions,

    R. Elhabyan, W. Shi, and M. St-Hilaire, “Coverage protocols for wireless sensor networks: Review and future directions,” Journal of Communications and Networks , vol. 21, no. 1, pp. 45–60, 2019

  8. [8]

    Distributed deployment algorithms for coverage improvement in a network of wireless mobile sensors: Relocation by virtual force,

    H. Mahboubi and A. G. Aghdam, “Distributed deployment algorithms for coverage improvement in a network of wireless mobile sensors: Relocation by virtual force,” IEEE Transactions on Control of Network Systems, vol. 4, no. 4, pp. 736–748, 2017

Show all 29 references
  1. [9]

    Coverage control for mobile sensing networks,

    J. Cortes, S. Martinez, T. Karatas, and F. Bullo, “Coverage control for mobile sensing networks,” IEEE Transactions on Robotics and Automation, vol. 20, no. 2, pp. 243–255, 2004

  2. [10]

    A distributed strategy to maximize coverage in a heterogeneous sensor network in the presence of obstacles,

    H. Mosalli and A. G. Aghdam, “A distributed strategy to maximize coverage in a heterogeneous sensor network in the presence of obstacles,” in Proceedings of 62nd IEEE Conference on Decision and Control, pp. 8426–8431, 2023

  3. [11]

    Distributed coordination of a network of nonidentical agents with limited commu- nication capabilities in the presence of fixed obstacles,

    H. Mahboubi, F. Sharifi, A. G. Aghdam, and Y . Zhang, “Distributed coordination of a network of nonidentical agents with limited commu- nication capabilities in the presence of fixed obstacles,” in Proceedings of American Control Conference , pp. 5398–5403, 2013

  4. [12]

    A gradient-based coverage optimization strategy for mobile sensor networks,

    J. Habibi, H. Mahboubi, and A. G. Aghdam, “A gradient-based coverage optimization strategy for mobile sensor networks,” IEEE Transactions on Control of Network Systems , vol. 4, no. 3, pp. 477– 488, 2017

  5. [13]

    Using occupancy grids for mobile robot perception and navigation,

    A. Elfes, “Using occupancy grids for mobile robot perception and navigation,” Computer, vol. 22, no. 6, pp. 46–57, 1989

  6. [14]

    A probabilistic coverage protocol for wireless sensor networks,

    M. Hefeeda and H. Ahmadi, “A probabilistic coverage protocol for wireless sensor networks,” in Proceedings of International Conference on Network Protocols , pp. 41–50, 2007

  7. [15]

    Coordinated locomo- tion and monitoring using autonomous mobile sensor nodes,

    S. Yoon, O. Soysal, M. Demirbas, and C. Qiao, “Coordinated locomo- tion and monitoring using autonomous mobile sensor nodes,” IEEE Transactions on Parallel and Distributed Systems , vol. 22, no. 10, pp. 1742–1756, 2011

  8. [16]

    Okabe, B

    A. Okabe, B. Boots, K. Sugihara, and S. N. Chiu, Spatial tessellations: concepts and applications of V oronoi diagrams. John Wiley and Sons, 2009

  9. [17]

    Sensor placement for effective cov- erage and surveillance in distributed sensor networks,

    S. S. Dhillon and K. Chakrabarty, “Sensor placement for effective cov- erage and surveillance in distributed sensor networks,” in Proceedings of IEEE Conference on Wireless Communications and Networking , vol. 3, pp. 1609–1614, 2003

  10. [18]

    Distributed deployment algorithms for improved cov- erage in a network of wireless mobile sensors,

    H. Mahboubi, K. Moezzi, A. G. Aghdam, K. Sayrafian-Pour, and V . Marbukh, “Distributed deployment algorithms for improved cov- erage in a network of wireless mobile sensors,” IEEE Transactions on Industrial Informatics , vol. 10, no. 1, pp. 163–174, 2014

  11. [19]

    Energy efficient coverage control in wireless sensor networks based on multi-objective genetic algorithm,

    J. Jia, J. Chen, G. Chang, and Z. Tan, “Energy efficient coverage control in wireless sensor networks based on multi-objective genetic algorithm,” Computers and Mathematics with Applications , vol. 57, no. 11-12, pp. 1756–1766, 2009

  12. [20]

    Collaborative allocation and optimization of path planning for static and mobile sensors in hybrid sensor networks for environment monitoring and anomaly search,

    Y . Guo, Z. Xu, and J. Saleh, “Collaborative allocation and optimization of path planning for static and mobile sensors in hybrid sensor networks for environment monitoring and anomaly search,” Sensors, vol. 21, no. 23, p. 7867, 2021

  13. [21]

    Relocation of mobile wireless sensors in the presence of obstacles,

    S. Mahfoudh, I. Khoufi, P. Minet, and A. Laouiti, “Relocation of mobile wireless sensors in the presence of obstacles,” in Proceeding of International Conference on Telecommunications , pp. 1–5, 2013

  14. [22]

    A distributed deterministic annealing algorithm for limited-range sensor coverage,

    A. Kwok and S. Martinez, “A distributed deterministic annealing algorithm for limited-range sensor coverage,” IEEE Transactions on Control Systems Technology , vol. 19, no. 4, pp. 792–804, 2010

  15. [23]

    A comprehensive review on techniques and challenges of energy harvesting from distributed renewable energy sources for wireless sensor networks,

    M. N. Hussain, M. A. Halim, M. Y . A. Khan, S. Ibrahim, and A. Haque, “A comprehensive review on techniques and challenges of energy harvesting from distributed renewable energy sources for wireless sensor networks,” Control Systems and Optimization Letters , vol. 2, no. 1, pp...

  16. [24]

    Deep rein- forcement learning for UA V-based SDWSN data collection,

    P. A. Karegar, D. Z. Al-Hamid, and P. H. J. Chong, “Deep rein- forcement learning for UA V-based SDWSN data collection,” Future Internet, vol. 16, no. 11, p. 398, 2024

  17. [25]

    A distributed reinforcement learning strategy to maximize coverage in a hybrid heterogeneous sensor network,

    H. Mosalli and A. G. Aghdam, “A distributed reinforcement learning strategy to maximize coverage in a hybrid heterogeneous sensor network,” in Proceedings of 63rd IEEE Conference on Decision and Control, pp. 2475–2480, 2024

  18. [26]

    Single and multi-agent deep reinforcement learning for ai-enabled wireless networks: A tutorial,

    A. Feriani and E. Hossain, “Single and multi-agent deep reinforcement learning for ai-enabled wireless networks: A tutorial,” IEEE Commu- nications Surveys and Tutorials , vol. 23, no. 2, pp. 1226–1252, 2021

  19. [27]

    Energy-efficient cluster-based routing protocol for heterogeneous wireless sensor net- work,

    P. Rawat, G. S. Rawat, H. Rawat, and S. Chauhan, “Energy-efficient cluster-based routing protocol for heterogeneous wireless sensor net- work,” Annals of Telecommunications , pp. 1–14, 2024

  20. [28]

    Coverage problems in sensor networks: A survey,

    B. Wang, “Coverage problems in sensor networks: A survey,” ACM Computing Surveys (CSUR) , vol. 43, no. 4, pp. 1–53, 2011

  21. [29]

    Derivatives of probability functions and some applica- tions,

    S. Uryasev, “Derivatives of probability functions and some applica- tions,” Annals of Operations Research , vol. 56, pp. 287–311, 1995

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.