Pith. sign in

REVIEW 4 major objections 4 minor 106 references

Emergence of the Traffic Autonomous Zone (TAZ) for Telecommunication Operations from Spatial Heterogeneity in Cellular Networks

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a single regionalization—the Traffic Autonomous Zone—can balance all major telecom performance indicators, and that in Taiyuan its MNCD-KE solutions reach the Pareto frontier while Louvain solutions sit at the corners.

desk verdict A useful framing for telecom regionalization, but the Pareto advantage over Louvain is largely manufactured by calibrating the method to scale bands that Louvain itself defines. read the letter →

arxiv 2501.09028 v1 pith:ZEMEDMET submitted 2025-01-11 cs.SI

Liyan Xu , Jintong Tang , Hezhishi Jiang , Hongbin Yu , Yihang Li , Qian Huang , Yinsheng Zhou , Jun Zhang
show 1 more author
Yu Liu
This is my paper · ORCID
classification cs.SI
keywords TrafficAutonomousZonetelecomregionalizationmulti-objectiveoptimizationcommunitydetectionnetworksspatialheterogeneityhumanmobilityParetofrontier
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

Telecom operations—capacity planning, handover optimization, and marketing—are driven by different physical quantities, and each spatial pattern is usually regionalized separately by ad hoc experience. The paper argues that these patterns all inherit clustered structure from human mobility, so a single regionalization can serve all of them. It formalizes the Traffic Autonomous Zone (TAZ) as a multi-objective optimization over five inputs (population, traffic, travel OD, proximity, and place semantics), and solves it with a multilevel community-detection plus kernel-extension algorithm (MNCD-KE). In a Taiyuan case study, TAZ solutions sit in the interior of the Pareto frontier across region counts, while Louvain solutions cluster at corners. If the claim holds, operators could use one deterministic city partition for planning, construction, maintenance, optimization, and marketing, a step toward autonomous networks.

What carries the argument

The load-bearing machinery is MNCD-KE (multilevel network community detection with kernel extension): basic spatial units (BSUs) based on road hierarchy and natural features are the nodes, semantic and quantity attributes are node features, and OD trips and BSU proximity are edge weights. Multilayer community detection assigns each BSU a membership vector across communities, and a depth-first kernel extension merges marginal BSUs into contiguous regions. Three objective functions drive the optimization: a semantics score combining intra-region entropy and inter-region semantic distance, Moran's I for population and traffic (subtracted from one), and modularity for the OD and proximity networks. Characteristic scales are calibrated by sweeping Louvain's resolution parameter and taking the stable region-count plateaus—fewer than 50, 50 to 100, 100 to 300, and more than 300—as the scales for evaluating TAZ partitions.

What would settle it

Re-run the Taiyuan comparison with MNCD-KE parameters fixed across all region-count groups (no calibration to Louvain plateaus) and check whether its solutions still form a balanced Pareto interior; if they fall to the corners or behind Louvain, the claimed multi-objective advantage collapses. Alternatively, apply the pipeline to a second city whose Louvain resolution sweep lacks stable plateaus; if no characteristic scales appear, the method has no principled basis for choosing the number of regions.

Watch

Extended reading notes

Core claim

The paper's central claim is that the spatial distributions of telecom performance indicators—population, call volume, traffic, handover flows, place semantics, and urban morphology—are not independent but all manifest clustered structures generated by the same human-mobility laws, so a unified regionalization is possible. It defines the Traffic Autonomous Zone (TAZ) as a partition that is both discovered (from these self-organizing clusters) and constructed (to satisfy operational constraints and trade-offs). The TAZ problem is cast as a multi-objective optimization with five objectives grouped into semantics, quantity, and interaction, and solved with MNCD-KE on basic spatial units. On Taiyuan data, the TAZ solutions are concentrated in the interior of the Pareto frontier across all region-count groups, whereas Louvain solutions sit mostly at the corners, and beyond 100 regions every Pareto solution comes from MNCD-KE. The paper presents this as evidence that TAZ balances the dimensions rather than optimizing any single one.

Load-bearing premise

The whole evaluation assumes that the stable region-count ranges found by sweeping Louvain's resolution parameter—fewer than 50, 50 to 100, 100 to 300, and more than 300—are genuine characteristic scales of Taiyuan's interaction networks, and that MNCD-KE can be calibrated to those same scales before being compared with Louvain.

Editorial extensions

If this is right

  • A single TAZ partition can be chosen from the Pareto frontier according to an operator's priority—population coverage, mobility stability, or high-value semantics—without re-solving the regionalization.
  • The same framework gives nested multilevel regions (district, sub-district, neighborhood, and community levels), so a city can be managed hierarchically.
  • TAZ identifies high-value service areas such as commercial hubs, universities, and industrial zones, enabling targeted service models such as streaming, gaming, and production-data transmission.
  • Replacing ad hoc micro-grid definitions with deterministic TAZ delineation makes telecom regionalization replicable across cities and across business units.
  • TAZ turns regionalization from an arbitrary administrative choice into a discovery of the city's self-organized structure, which is the basis for divide-and-rule autonomy.

Reading between the lines

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

  • A natural testable extension is to run the same pipeline in several cities and check whether the stable region-count plateaus recur at comparable administrative levels; if they do, the TAZ method becomes a general tool rather than a Taiyuan-specific fit.
  • The comparison with Louvain is the strongest evidence, but a stricter test would compare TAZ against other multi-objective regionalizers, such as max-p compact regions, and test sensitivity to the choice of BSU boundaries and to the semantic-entropy weighting.
  • Because the semantic objective is self-defined, an external validation against planned land-use maps or administrative function labels would clarify whether TAZ boundaries match ground-truth places.
  • If TAZ really captures stable interaction communities, mobility management could optimize handover parameters at the zone level to reduce unnecessary handovers; this is a concrete operational consequence the paper leaves implicit.
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. The paper proposes a 'Traffic Autonomous Zone (TAZ)' as a unified regionalization scheme for telecommunication operations, formulated as a multi-objective optimization problem balancing semantic, population/traffic, and interaction (OD and proximity) criteria. The authors use their previously developed MNCD-KE algorithm (multi-layer network community detection with kernel extension) to solve the problem and apply it to a case study of central Taiyuan, China. They compare the resulting partitions against Louvain community detection across a Pareto-frontier analysis, claiming that MNCD-KE solutions occupy the middle of the frontier and thus achieve a better balance among the three objectives.

Significance. If substantiated, the TAZ concept would offer a principled alternative to ad hoc telecom service regionalization, with clear relevance to autonomous network operations and urban management. The paper's strength is its conceptual synthesis of geographic regionalization theory (attributes vs. interactions, characteristic scales, MAUP) with a concrete algorithmic pipeline, and its attempt to evaluate a multi-objective solution against a standard community-detection baseline. The framing is valuable and the underlying problem is well motivated. However, the empirical evidence is currently limited to a single city and relies on self-defined metrics, and the comparison with Louvain is complicated by a partially circular use of Louvain-derived characteristic scales. The central claim of Pareto superiority therefore needs substantial additional support.

major comments (4)
  1. [§3.2.2, §3.3, §4.3.1] The comparison between MNCD-KE and Louvain is partially circular in its use of characteristic scales. Section 3.3 states that the characteristic scale ranges (fewer than 50, 50–100, 100–300, more than 300 regions) are detected by sweeping Louvain's resolution parameter on the travel OD and BSU distance networks. Section 3.2.2 then says that MNCD-KE parameters are 'adjusted according to the characteristic scale to ensure that the regionalization results are comparable to the characteristic scale.' These Louvain-derived bands are subsequently used in Section 4.3.1 to group both methods' solutions before drawing the Pareto-frontier comparison. This means Louvain is evaluated on a region-count grid that is itself a byproduct of Louvain's stability plateaus, while MNCD-KE is given the freedom to search within those same bands. The conclusion that the region-number ranges are 'reasonable for our method' and that MNCD-KE achieves a superior Pareto balance is therefore not a method-level comparison but a comparison conditional on a Louvain-influenced parameterization. I recommend either detecting characteristic scales independently of the baseline algorithm (e.g., using statistical scale-detection methods on the spatial processes themselves) or, at minimum, reporting results at equal region counts using a common set of externally justified scales.
  2. [§4.2.1, Eq. (7)] The semantic objective function in Eq. (7) is insufficiently specified and likely malformed. The notation is inconsistent: the region index j is used both for the target region and in the denominator of the outer sum; the semantic category index t appears both as a running index and as the total number of categories; the log terms are written without an explicit base or a well-defined argument, and the negative sign placement is ambiguous. Because semantics is one of the three consolidated indicators in the Pareto-frontier claim, an ill-defined objective function prevents readers from verifying what is actually optimized and evaluated. The authors should present a clean equation with all variables explicitly defined, or replace it with a standard measure (e.g., area-weighted Shannon entropy with a clear normalization) and state how the balancing with inter-regional dissimilarity is performed.
  3. [§4.2.3, Eqs. (11)–(14)] The interaction objectives are defined as modularity Q of the network of BSUs with OD or proximity as edge weights, evaluated on the regionalization Z. But MNCD-KE itself performs community detection on a multi-layer network that includes these same interaction data (per Sections 3.2.1 and 3.2.2), so the evaluation metric overlaps with the optimization machinery of MNCD-KE. Louvain, in contrast, directly maximizes modularity on a single interaction layer; evaluating both methods on modularity does not establish that MNCD-KE finds better interaction structure, but rather that, after embedding the same interaction data in its multi-layer objective, MNCD-KE produces partitions with higher modularity on that layer. To make the comparison fair, the authors should specify the full objective functions actually optimized by MNCD-KE, and then evaluate both methods on those objectives as well as on each individual metric separately, rather than using one of the optimized terms as the common evaluator.
  4. [§4.3.1] The Pareto-frontier analysis is based entirely on the self-defined objective functions (semantics, quantity, interaction) and is not validated against any external benchmark or alternative regionalization method. The claim that MNCD-KE solutions 'are mostly located in the middle of the Pareto frontier' while Louvain solutions 'are mostly at the corner' is a descriptive observation of a single city's data, with no indication of sensitivity to parameter choices, no error bars, and no comparison with, e.g., administrative boundaries, random contiguous partitions, or other regionalization baselines (such as max-p regions or AZP). Moreover, the evaluation uses only Taiyuan, so the generalizability of the TAZ framework to other cities is unsupported. To support the paper's central claim, I recommend adding at least one additional city, reporting the distribution of objective values across parameter configurations, and comparing against one or more non-community-detection baselines.
minor comments (4)
  1. [§3.1, Eqs. (1)–(6)] The problem statement contains notational inconsistencies that should be fixed: in Eq. (5), the index p on the right-hand side should be r (the text refers to 'the p-th regionalization unit' but uses r elsewhere); in Eqs. (8) and (11), the indices in the summation (m and a; α and β) are not consistently defined, and the expected-degree term in Eq. (11) lacks the standard 2m normalization. These issues make the formal model harder to follow.
  2. [§4.2.2] The use of Moran's I as a regionalization objective is justified in the text as 'the closer the population and traffic results are to spatial negative correlation, the better,' but this is an assumption that may not hold for all cities or all regionalization goals. It would be helpful to discuss why maximizing negative spatial autocorrelation is an appropriate objective for telecom operations, and whether this was validated in any prior work.
  3. [§4.3.4, Table 2] Table 2 reports objective function values for three selected schemes (Z1, Z2, Z3), but there is no indication of the range of values across the Pareto front or the sensitivity of these schemes to small parameter changes. Reporting only the selected schemes makes it difficult to assess how robust the scenario-specific choices are.
  4. [General] The MNCD-KE algorithm is only described briefly and relies heavily on the authors' prior publication [74]. To make this paper self-contained, I recommend adding a pseudo-code or more detailed algorithmic description in an appendix, especially for the kernel-extension step that is central to the method.

Circularity Check

2 steps flagged · score 6.0 of 10

Pareto advantage is partly manufactured: Louvain defines the characteristic-scale bands, MNCD-KE is tuned into those bands, and the same objective functions used for optimization are used as the evaluation metrics.

  1. fitted input called prediction [Sections 3.2.2, 3.3, 4.3.1]
    "Adjust the parameters according to the characteristic scale to ensure that the regionalization results are comparable to the characteristic scale. [...] we use the Louvain algorithm with a resolution parameter ... to determine the characteristic scales. [...] Firstly, our algorithm's position in the Pareto space is relatively concentrated, which indicates that the region number ranges based on the Louvain algorithm are also reasonable for our method."

    The characteristic scales (<50, 50-100, 100-300, >300 regions) are detected by sweeping Louvain's resolution parameter on the OD and BSU-distance interaction networks (Section 3.3). Section 3.2.2 then states that MNCD-KE parameters are adjusted according to those characteristic scales to make the regionalization results comparable to them. Consequently, MNCD-KE solutions are intentionally tuned to fall into Louvain-derived region-count bands. When Section 4.3.1 subsequently groups the Pareto solutions by these same bands and cites their concentration as evidence that the Louvain-based ranges 'are also reasonable for our method,' it is using the calibration target as the validation.

  2. other [Section 4.2, 4.3.1]
    "These objective functions also serve as evaluation metrics for the regionalization solutions."

    The five objective functions defined in Section 4.2 (semantic entropy/distance, Moran's I for population and traffic, and modularity for OD and proximity) are exactly the indicators used to compare MNCD-KE against Louvain and to construct the Pareto frontier in Section 4.3.1. MNCD-KE's parameter search is designed to optimize these functions, whereas the Louvain baseline optimizes modularity alone. The headline finding that MNCD-KE solutions occupy the middle of the Pareto frontier while Louvain solutions sit at corners is therefore partly a consequence of evaluating both methods with MNCD-KE's own objective functions.

full rationale

The clearest circular step is the characteristic-scale calibration. The paper first uses Louvain to identify stable region-count ranges, then adjusts MNCD-KE parameters so that its results are 'comparable to the characteristic scale,' and finally treats the concentration of MNCD-KE solutions within those ranges as evidence that Louvain-derived ranges are reasonable for the new method. That is a fitted input being presented as confirmation: the validation target is the calibration target. A second, milder self-referentiality comes from Section 4.2, where the same functions used as optimization objectives are also used as evaluation metrics; since Louvain only optimizes modularity, the Pareto-frontier advantage of a multi-objective method on its own objectives is substantially built into the comparison. The paper's reliance on the authors' prior MNCD-KE framework [74] is not by itself circular, because that framework is a separately published method, but the present evaluation provides no external benchmark against which either method is independently validated. The TAZ application is a real case study with genuine data, and the multi-objective formulation is not vacuous, so the paper is not wholly circular; however, the central comparative claim is partially forced by construction, warranting a score of 6 rather than a higher or lower value.

Assumptions & free parameters 4 free parameters · 5 assumptions · 2 invented entities

The paper rests on four classes of assumptions: (1) the geographic consistency of telecom objectives, argued qualitatively from the literature; (2) the reliability of the BSU street-block layer from an unreleased dataset; (3) the legitimacy of using Louvain-derived characteristic scales as the calibration target; and (4) the validity of the self-defined objective functions, including the preference for negative spatial autocorrelation. The only genuinely new entity is the TAZ label; it carries no independent observable prediction.

free parameters (4)
  • Characteristic scale boundaries = 50, 100, 300 regions
    Used to name four scale levels and group results; identified from Louvain resolution sweeps (Section 4.3.2) and then used to calibrate the proposed method.
  • Area constraints A_min and A_max = not stated
    Constraint (6) filters regionalization units by area; thresholds are not given, affecting which BSUs are excluded from the analysis.
  • MNCD-KE internal parameters (membership threshold, kernel extension thresholds) = not stated
    Section 3.2.2 says 'Adjust the parameters according to the characteristic scale', so parameter values are tuned to data rather than published.
  • Semantic entropy base and normalization choices = not stated
    Equation (7) uses entropy and normalized semantic vectors; the log base and any normalization details are unspecified.
assumptions (5)
  • domain assumption Human mobility and telecommunication activity distributions share universal spatial scaling laws, implying consistency among regionalization objectives.
    Section 2.2 argues for this consistency via literature; it is the conceptual basis for TAZ and is not derived in this paper.
  • domain assumption Spatial interaction communities are geographically contiguous due to distance decay.
    Sections 2.1.2 and 2.2 rely on this property to justify functional regionalization, citing refs. [49,50,51].
  • domain assumption The BSU layer derived from street-block divisions (MSDCW) is a valid operational unit.
    Section 3.1.2.2 relies on BSUs from ref. [88], a dataset under review and not available for inspection.
  • ad hoc to paper Louvain resolution sweeps reveal the true characteristic scales of the region-generation process.
    Section 4.3.2 uses stable ranges of Louvain community counts to define scales; this imports Louvain's structure into the proposed method's calibration.
  • ad hoc to paper Negative spatial autocorrelation (Moran's I close to -1) indicates better regionalization for population and traffic.
    Section 4.2.2 states this preference without external justification or sensitivity analysis.
invented entities (2)
  • Traffic Autonomous Zone (TAZ)
    purpose: A named regionalization scheme intended to unify telecom operational and service zones.
    TAZ is a conceptual label for the output of a regionalization algorithm; no independent observable is predicted beyond the city partitions shown in the case study.
  • Valuable service regions
    purpose: Zones identified as needing distinct telecom services (e.g., commercial hub, university, factory).
    These are selected examples without quantitative ground-truth comparison; their identification is not assigned a measurable external target.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Emergence of the Traffic Autonomous Zone (TAZ) for Telecommunication Operations from Spatial Heterogeneity in Cellular Networks." pith.science (2026). https://pith.science/paper/ZEMEDMET

@misc{pith2026250109028,
  author       = {Pith},
  title        = {Pith review of: Emergence of the Traffic Autonomous Zone (TAZ) for Telecommunication Operations from Spatial Heterogeneity in Cellular Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZEMEDMET}},
  note         = {Machine review of arXiv:2501.09028}
}
read the original abstract

In the field of telecommunications, various operations are driven by different physical quantities. Each has its own patterns in time and space, but all show some clustered structures in their spatial distribution. This reflects a unified rule of human mobility, suggesting the consistency among different telecommunication regionalization objectives. With this in mind, regionalization can be used to identify these patterns and can be applied to improve management efficiency in the context of "autonomous networks". This article introduces the "Traffic Autonomous Zone (TAZ)" concept. This approach aims to create a reasonable unified regionalization scheme by identifying spatial clusters. It is not just a practical way to partition cities based on telecommunications needs, but it also captures self-organization structure of cities in essence. We present examples of this regionalization method using real data. Compared to the popular Louvain community detection method, our approach is on the Pareto frontier, allowing for a balance among various metrics in telecommunications.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

106 extracted references · 80 canonical work pages

  1. [1]

    College of Architecture and Landscape Architecture, Peking University, Beijing, China

  2. [2]

    Institute of Remote Sensing and Geographical Information Systems, School of Earth and Space Sciences, Peking University, Beijing, China

  3. [3]

    Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, China

  4. [4]

    Global Technical Service Dept, Huawei Technologies, Beijing, China

  5. [5]

    autonomous networks

    Global Technical Service Dept, Huawei Technologies, Shanghai, China. *: Corresponding Author. 2 Abstract In the field of telecommunications, various operations are driven by different physical quantities. Each has its own patterns in time and space, but all show some clustered structures in their spatial distribution. This reflects a unified rule of human...

  6. [6]

    It should reflect the inherent self -organizing spatial structure of key performance indicators involved in telecommunications such, rather than being an arbitrary desi gnation. Particularly, it must consider a series of KPIs and KQIs in telecommunication operations, including user capacity within the region, interactions among regions, urban place semant...

  7. [7]

    This means that, in addition to the aforementioned KPIs and KQIs, the main indicators of the above services should also be included in the regionalization objectives

    It should encompass the full process of various telecommunication operations , such as network planning, construction, maintenance, optimization, and marketing. This means that, in addition to the aforementioned KPIs and KQIs, the main indicators of the above services should also be included in the regionalization objectives

  8. [8]

    first-order

    It should have consistent regionalization logic and determinate regionalization results, while allowing for necessary operational flexibility. This requires that the TAZ, while considering the 5 needs of the various types of operations mentioned above, allows for prioritization and trade-offs among multiple potential objectives. However, the final spatial...

    2011

Show all 106 references
  1. [9]

    Directly use the regionalization results of the MNCD-KE algorithm, aggregate each region into a multipolygon as a new regionalization unit, then regionalize it once again with MNCD-KE, aggregating into larger regions to achieve multi-layer nesting

  2. [10]

    Regard higher -level BSU as higher levels, and the lower-level BSU contained within them as lower levels

    Directly use the nested relationship of BSU . Regard higher -level BSU as higher levels, and the lower-level BSU contained within them as lower levels. The reason for this method is that BSU itself is determined based on city road network levels, and road networks naturally pa...

  3. [11]

    spatial attributes

    Obtain regionalization schemes separately based on each characteristic scale. Note that the division results of this method are not fully nested from top to bottom. Next, we use the Louvain algorithm with a resolution parameter for community detection on the inputted interacti...

  4. [12]

    This implies a preference for regionalizati on based on population distribution

    User coverage. This implies a preference for regionalizati on based on population distribution. Therefore, we selected a partitioning scheme with a higher population indicator, corresponding to the scheme 𝑍𝑍1 in Fig. 6(a)

  5. [13]

    This means as few switching among cells and base stations as possible, which actually requires a schema that emphasizes OD interactions

    Crowd mobility coverage. This means as few switching among cells and base stations as possible, which actually requires a schema that emphasizes OD interactions. Therefore, we selected 𝑍𝑍2 (Fig. 6(b)), a schema with high interaction metrics

  6. [14]

    discover

    High-value area coverage. This suggests that areas with special semantics should be delineated as much as possible, and therefore regionalization schema 𝑍𝑍3 with high semantic metrics was chosen (Fig. 6(c)). Fig. 6 Regionalization schemas for each scenario. (a) 𝒁𝒁𝟏𝟏; (b) 𝒁𝒁𝟐𝟐;...

  7. [15]

    Autonomous networks: Empowering digital transformation for the telecoms industry,

    A. R. E. Boasman-Patel et al., “Autonomous networks: Empowering digital transformation for the telecoms industry,” May 2019

    2019

  8. [16]

    Striding Towards the Intelligent World White Paper 2023 ADN: Intelligence + Connectivity for High Autonomous Networks,

    Huawei Technologies, “Striding Towards the Intelligent World White Paper 2023 ADN: Intelligence + Connectivity for High Autonomous Networks,” 2023

    2023

  9. [17]

    Technical Specification Group Services and System Aspects; Release 17 Description; Summary of Rel-17 Work Items (3GPP TR 21.917 Release 17),

    3GPP, “Technical Specification Group Services and System Aspects; Release 17 Description; Summary of Rel-17 Work Items (3GPP TR 21.917 Release 17),” 2022

    2022

  10. [18]

    A quantitative comparison of regionalization methods,

    O. Aydin, M. V. Janikas, R. M. Assunção, and T. H. Lee, “A quantitative comparison of regionalization methods,” Int. J. Geogr. Inf. Sci., vol. 35, no. 11, pp. 2287–2315, 2021

    2021

  11. [19]

    Space-time relationship between urban municipal district adjustment and built-up area expansion in China,

    K. Wang and W. Qi, “Space-time relationship between urban municipal district adjustment and built-up area expansion in China,” Chinese Geogr. Sci., vol. 27, no. 2, pp. 165–175, May 2017

    2017

  12. [20]

    Constrained clustering via diagrams: A unified theory and its application to electoral district design,

    A. Brieden, P. Gritzmann, and F. Klemm, “Constrained clustering via diagrams: A unified theory and its application to electoral district design,” Eur. J. Oper. Res., vol. 263, no. 1, pp. 18–34, Nov. 2017

    2017

  13. [21]

    The zoning map and American city form,

    S. T. Moga, “The zoning map and American city form,” J. Plan. Educ. Res., vol. 37, no. 3, pp. 271–285, Sep. 2017

    2017

  14. [22]

    Research on microgrid planning method and evolution strategy,

    L. Chen, W. Fang, G. Cheng, and S. Wang, “Research on microgrid planning method and evolution strategy,” Dianxin Jishu/Telecom Technol., vol. 10, pp. 18–23, 2017

    2017

  15. [23]

    An example analysis of the microgrid construction in the comprehens ive business area,

    F. Meng and F. Wang, “An example analysis of the microgrid construction in the comprehens ive business area,” Tongxun Shijie/Communication World, vol. 6, pp. 15–16, 2017

    2017

  16. [24]

    The modifiable areal unit problem,

    S. Openshaw, “The modifiable areal unit problem,” in Quantitative geography: A British view , Routledge & Kegan Paul, 1981, pp. 60–69

    1981

  17. [25]

    The validity and usefulness of laws in geographic information science and geography,

    M. F. Goodchild, “The validity and usefulness of laws in geographic information science and geography,” Ann. Assoc. Am. Geogr., vol. 94, no. 2, pp. 300–303, 2004

    2004

  18. [26]

    Wicked Problems,

    C. W. Churchman, “Wicked Problems,” Manage. Sci., vol. 14, no. 4, p. B-141-B-142, Dec. 1967

    1967

  19. [27]

    Geography and the Gerrymander,

    C. O. Sauer, “Geography and the Gerrymander,” Am. Polit. Sci. Rev., vol. 12, no. 3, pp. 403–426, Aug. 23 1918

    1918

  20. [28]

    The scales of human mobility,

    L. Alessandretti, U. Aslak, and S. Lehmann, “The scales of human mobility,” Nature , vol. 587, no. November, pp. 402–407, 2020

    2020

  21. [29]

    Modelling the scaling properties of human mobility,

    C. Song, T. Koren, P. Wang, and A. L. Barabási, “Modelling the scaling properties of human mobility,” Nat. Phys., vol. 6, no. 10, pp. 818–823, 2010

    2010

  22. [30]

    Limits of predictability in human mobility,

    C. Song, Z. Qu, N. Blumm, and A.-L. Barabási, “Limits of predictability in human mobility,” Science (80-. )., vol. 327, no. 5968, pp. 1018–1021, Feb. 2010

    2010

  23. [31]

    Understanding individual human mobility patterns,

    M. C. González, C. A. Hidalgo, and A. L. Barabási, “Understanding individual human mobility patterns,” Nature, vol. 453, pp. 779–782, 2008

    2008

  24. [32]

    The universal visitation law of human mobility,

    M. Schläpfer et al., “The universal visitation law of human mobility,” Nature, vol. 593, pp. 522–527, May 2021

    2021

  25. [33]

    The statistical physics of cities,

    M. Barthelemy, “The statistical physics of cities,” Nat. Rev. Phys., vol. 1, no. 6, pp. 406–415, Jun. 2019

    2019

  26. [34]

    A unified theory of urban living,

    L. Bettencourt and G. West, “A unified theory of urban living,” Nature, vol. 467, no. 7318, pp. 912– 913, 2010

    2010

  27. [35]

    Radio propagation path loss models for 5G cellular networks in the 28 GHZ and 38 GHZ millimeter -wave bands,

    A. Sulyman, A. Nassar, M. Samimi, G. Maccartney, T. Rappaport, and A. Alsanie, “Radio propagation path loss models for 5G cellular networks in the 28 GHZ and 38 GHZ millimeter -wave bands,” IEEE Commun. Mag., vol. 52, no. 9, pp. 78–86, Sep. 2014

    2014

  28. [36]

    Planning of 2G Networks,

    J. T. J. Penttinen, “Planning of 2G Networks,” in The Telecommunications Handbook , John Wiley & Sons, Ltd, 2015, pp. 669–742

    2015

  29. [37]

    A computer movie simulating urban growth in the Detroit region,

    W. R. Tobler, “A computer movie simulating urban growth in the Detroit region,” Econ. Geogr. , vol. 46, pp. 234–240, Jun. 1970

    1970

  30. [38]

    Anselin, Spatial econometrics: Methods and models

    L. Anselin, Spatial econometrics: Methods and models. Dordrecht: Springer Netherlands, 1988

    1988

  31. [39]

    Characterizing user activity in WiFi networks: University campus and urban area case studies,

    L. Oliveira, K. Obraczka, and A. Rodríguez, “Characterizing user activity in WiFi networks: University campus and urban area case studies,” in MSWiM 2016 - Proceedings of the 19th ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems ...

    2016

  32. [40]

    Discrete models and algorithms for the cap acitated location problems arising in UMTS network planning,

    E. Amaldi, A. Capone, and F. Malucelli, “Discrete models and algorithms for the cap acitated location problems arising in UMTS network planning,” in Proceedings of the 5th International Workshop on Discrete Algorithms and Methods for Mobile Computing and Communications , 2001, pp. 1–8

    2001

  33. [41]

    Scenarios, requirements and KPIs for 5G mobile and wireless system,

    P. Popovski et al., “Scenarios, requirements and KPIs for 5G mobile and wireless system,” May 2013

    2013

  34. [42]

    End-to-End wireless network performance evaluation based on application quality indicators,

    D. Chheda et al., “End-to-End wireless network performance evaluation based on application quality indicators,” in 2018 IEEE Global Communications Conference, 2018, pp. 1–6

    2018

  35. [43]

    QoS KPI and QoE KQI relationship for LTE video streaming and VoLTE services,

    M. Vaser and S. Forconi, “QoS KPI and QoE KQI relationship for LTE video streaming and VoLTE services,” in 2015 9th International Conference on Next Generation Mobile Applications, Services and Technologies, 2016, pp. 318–323

    2015

  36. [44]

    WCDMA KPI framework definition methods and applications,

    X. Ke, W. Li, L. Rui, X. Qiu, and S. Guo, “WCDMA KPI framework definition methods and applications,” in 2010 International Conference on Computer Engineering and Technology, 2010, vol. 4, pp. 471–475

    2010

  37. [45]

    Radio frequency optimization & QoS evaluation in operational GSM network,

    B. Haider, M. Zafrullah, and M. K. Islam, “Radio frequency optimization & QoS evaluation in operational GSM network,” in Proceedings of the World Congress on Engineeri ng and Computer Science 2009, 2009

    2009

  38. [46]

    Analysis of service transaction flow based on user’s actions to develop KQIs for WiBro service,

    D. W. Kim, H. M. Lim, J. H. Yoo, and S. H. Kim, “Analysis of service transaction flow based on user’s actions to develop KQIs for WiBro service,” in Fourth Advanced International Conference on Telecommunications, 2008, pp. 77–84

    2008

  39. [47]

    Telecom network and service management: An operator survey,

    S. Wallin and V. Leijon, “Telecom network and service management: An operator survey,” in 24 IFIP/IEEE International Conference on Management of Multimedia Networks and Services (WMNS 2009), 2009, vol. 5842 LNCS, pp. 15–26

    2009

  40. [48]

    Analysis of failure and recovery rates in a wireless telecommunications system,

    S. M. Matz, L. G. Votta, and M. Malkawi, “Analysis of failure and recovery rates in a wireless telecommunications system,” in Proceedings of the 2002 International Conference on Dependable Systems and Networks, 2002, pp. 687–693

    2002

  41. [49]

    Coverage and Capacity Planning of 4G Networks,

    A. Elnashar, “Coverage and Capacity Planning of 4G Networks,” in Design, Deployment and Performance of 4G-LTE Networks: A Practical Approach, Wiley Blackwell, 2014, pp. 349–444

    2014

  42. [50]

    Urban population densities,

    C. Clark, “Urban population densities,” J. R. Stat. Soc. Ser. A, vol. 114, no. 4, pp. 490 –496, Jul. 1951

    1951

  43. [51]

    The spatial variation of urban population densities,

    B. E. Newling, “The spatial variation of urban population densities,” Geogr. Rev. , vol. 59, no. 2, pp. 242–252, Apr. 1969

    1969

  44. [52]

    The distribution of population and employment in a polycentric city: The case of Los Angeles,

    P. Gordon, H. W. Richardson, and H. L. Wong, “The distribution of population and employment in a polycentric city: The case of Los Angeles,” Environ. Plan. A, vol. 18, no. 2, pp. 161–173, 1986

    1986

  45. [53]

    Population and employment densities: Structure and change,

    K. A. Small and S. Song, “Population and employment densities: Structure and change,” J. Urban Econ., vol. 36, no. 3, pp. 292–313, Nov. 1994

    1994

  46. [54]

    From mobile phone data to the spatial structure of cities,

    T. Louail et al., “From mobile phone data to the spatial structure of cities,” Sci. Rep., vol. 4, p. 5276, 2014

    2014

  47. [55]

    On the spatial distribution of base stations and its relation to the traffic density in cellular networks,

    S. Zhou, D. Lee, B. Leng, X. Zhou, H. Zhang, and Z. Niu, “On the spatial distribution of base stations and its relation to the traffic density in cellular networks,” IEEE Access, vol. 3, pp. 998–1010, 2015

    2015

  48. [56]

    Understanding traffic dynamics in cellular data networks,

    U. Paul, A. P. Subramanian, M. M. Buddhikot, and S. R. Das, “Understanding traffic dynamics in cellular data networks,” in 2011 Proceedings IEEE INFOCOM, 2011, pp. 882–890

    2011

  49. [57]

    Spatial modeling of the traffic density in cellular networks,

    D. Lee, S. Zhou, X. Zhong, Z. Niu, X. Zhou, and H. Zhang, “Spatial modeling of the traffic density in cellular networks,” IEEE Wirel. Commun., vol. 21, no. 1, pp. 80–88, 2014

    2014

  50. [58]

    The origins of scaling in cities,

    L. M. A. Bettencourt, “The origins of scaling in cities,” Science (80-. )., vol. 340, no. 6139, pp. 1438– 1441, 2013

    2013

  51. [59]

    Challenges of using mobile phone signalling data to estimate urban population density: Towards smart cities and sustainable urban development,

    J. Yang, Y. Shi, C. Yu, and S. J. Cao, “Challenges of using mobile phone signalling data to estimate urban population density: Towards smart cities and sustainable urban development,” Indoor Built Environ., vol. 29, no. 2, pp. 147–150, Feb. 2020

    2020

  52. [60]

    Analytical methods and applications of spatial interactions in the era of big data,

    Y. Liu et al., “Analytical methods and applications of spatial interactions in the era of big data,” Dili Xuebao/Acta Geogr. Sin., vol. 75, pp. 1523–1538, 2020

    2020

  53. [61]

    A trade-off between unnecessary handover and handover failure for heterogeneous networks,

    M. Alhabo, L. Zhang, and N. Nawaz, “A trade-off between unnecessary handover and handover failure for heterogeneous networks,” in European Wireless 2017, 2017, pp. 167–172

    2017

  54. [62]

    Understanding the patterns of car travel,

    L. Pappalardo, S. Rinzivillo, Z. Qu, D. Pedreschi, and F. Giannotti, “Understanding the patterns of car travel,” Eur. Phys. J. Spec. Top., vol. 215, pp. 61–73, 2013

    2013

  55. [63]

    Redrawing the map of Great Britain from a network of human interact ions,

    C. Ratti et al., “Redrawing the map of Great Britain from a network of human interact ions,” PLoS One, vol. 5, no. 12, p. e14248, 2010

    2010

  56. [64]

    Discovering spatial interaction communities from mobile phone data,

    S. Gao, Y. Liu, Y. Wang, and X. Ma, “Discovering spatial interaction communities from mobile phone data,” Trans. GIS, vol. 17, no. 3, pp. 463–481, 2013

    2013

  57. [65]

    Spatial interaction modelling,

    J. R. Roy and J. C. Thill, “Spatial interaction modelling,” Pap. Reg. Sci., vol. 83, no. 1, pp. 339–361, 2004

    2004

  58. [66]

    Delineating a hierarchical organization of ranked urban clusters using a spatial interaction network,

    T. Jia, X. Luo, and X. Li, “Delineating a hierarchical organization of ranked urban clusters using a spatial interaction network,” Comput. Environ. Urban Syst., vol. 87, p. 101617, May 2021

    2021

  59. [67]

    Tuan, Space and place: The perspective of experience

    Y.-F. Tuan, Space and place: The perspective of experience. Minneapolis: University of Minnesota Press, 1977

    1977

  60. [68]

    Corbusier, The Athens Charter

    L. Corbusier, The Athens Charter. New York: Grossman Publishers, 1973. 25

    1973

  61. [69]

    Study on Scenarios and Requirements for Next Generation Access Technologies (3GPP TR 38.913 version 14.2.0 Release 14),

    European Telecommunications Standards Institute, “Study on Scenarios and Requirements for Next Generation Access Technologies (3GPP TR 38.913 version 14.2.0 Release 14),” 2017

    2017

  62. [70]

    Formalizing place in geographic information systems,

    M. F. Goodchild, “Formalizing place in geographic information systems,” in Communities, Neighborhoods, and Health, Springer, New York, NY, 2011, pp. 21–33

    2011

  63. [71]

    Extracting urban functional regions from points of interest and human activities on location-based social networks,

    S. Gao, K. Janowicz, and H. Couclelis, “Extracting urban functional regions from points of interest and human activities on location-based social networks,” Trans. GIS, vol. 21, no. 3, pp. 446–467, 2017

    2017

  64. [72]

    Social sensing: A new approach to understanding our socioeconomic environments,

    Y. Liu et al., “Social sensing: A new approach to understanding our socioeconomic environments,” Ann. Assoc. Am. Geogr., vol. 105, no. 3, pp. 512–530, 2015

    2015

  65. [73]

    Inferring ‘high-frequent’ mixed urban functions from telecom traffic,

    J. Tang et al., “Inferring ‘high-frequent’ mixed urban functions from telecom traffic,” Environ. Plan. B Urban Anal. City Sci., vol. 51, no. 8, pp. 1775–1793, Dec. 2024

    2024

  66. [74]

    A typology of street patterns,

    R. Louf and M. Barthélemy, “A typology of street patterns,” J. R. Soc. Interface, vol. 11, no. 101, p. 20140924, Dec. 2014

    2014

  67. [75]

    Identifying borders of activity spaces and quantifying border effects on intra-urban travel through spatial interaction network,

    M. Jin, L. Gong, Y. Cao, P. Zhang, Y. Gong, and Y. Liu, “Identifying borders of activity spaces and quantifying border effects on intra-urban travel through spatial interaction network,” Comput. Environ. Urban Syst., vol. 87, p. 101625, May 2021

    2021

  68. [76]

    The P1 P2/D Hypothesis: On the Intercity Movement of Persons,

    G. K. Zipf, “The P1 P2/D Hypothesis: On the Intercity Movement of Persons,” Am. Sociol. Rev., vol. 11, no. 6, pp. 677–686, Dec. 1946

    1946

  69. [77]

    Intervening opportunities: A theory relating mobility and distance,

    S. A. Stouffer, “Intervening opportunities: A theory relating mobility and distance,” Am. Sociol. Rev. , vol. 5, no. 6, p. 845, Dec. 1940

    1940

  70. [78]

    Inferring spatial interaction patterns from sequential snapshots of spatial distributions,

    D. Zhu, Z. Huang, L. Shi, L. Wu, and Y. Liu, “Inferring spatial interaction patterns from sequential snapshots of spatial distributions,” Int. J. Geogr. Inf. Sci., vol. 32, no. 4, pp. 783 –805, 2018

    2018

  71. [79]

    Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting,

    B. Yu, H. Yin, and Z. Zhu, “Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting,” in IJCAI International Joint Conference on Artificial Intelligence , 2018, vol. 2018-July, pp. 3634–3640

    2018

  72. [80]

    Region2Vec: community detection on spatial networks using graph embedding with node attributes and spatial interactions,

    Y. Liang, J. Zhu, W. Ye, and S. Gao, “Region2Vec: community detection on spatial networks using graph embedding with node attributes and spatial interactions,” in Proceedings of the 30th International Conference on Advances in Geographic Information Systems, 2022, pp. 1–4

    2022

  73. [81]

    Enhancing Urban Resilience through Spatial Interaction- based City Management Zoning,

    H. Jiang, L. Xu, J. Li, J. Liu, and Y. Shen, “Enhancing Urban Resilience through Spatial Interaction- based City Management Zoning,” Ann. Am. Assoc. Geogr., vol. 114, no. 6, pp. 1310–1329, 2024

    2024

  74. [82]

    A Deep Gravity model for mobility flows generation,

    F. Simini, G. Barlacchi, M. Luca, and L. Pappalardo, “A Deep Gravity model for mobility flows generation,” Nat. Commun., vol. 12, no. 1, Dec. 2021

    2021

  75. [83]

    Flow trace: A novel representation of intra - urban movement dynamics,

    X. Xing, Y. Yuan, Z. Huang, X. Peng, P. Zhao, and Y. Liu, “Flow trace: A novel representation of intra - urban movement dynamics,” Comput. Environ. Urban Syst., vol. 96, p. 101832, Aug. 2022

    2022

  76. [84]

    Re-examining urban region and inferring regional function based on spatial–temporal interaction,

    H. Tao, K. Wang, L. Zhuo, and X. Li, “Re-examining urban region and inferring regional function based on spatial–temporal interaction,” Int. J. Digit. Earth, vol. 12, no. 3, pp. 293–310, Mar. 2019

    2019

  77. [85]

    Emergence of hierarchy in cost-driven growth of spatial networks,

    R. Louf, P. Jensen, and M. Barthelemy, “Emergence of hierarchy in cost-driven growth of spatial networks,” Proc. Natl. Acad. Sci. U. S. A., vol. 110, no. 22, pp. 8824–8829, May 2013

    2013

  78. [86]

    Charter of the New Urbanism,

    Congress for the New Urbanism, “Charter of the New Urbanism,” Bull. Sci. Technol. Soc., vol. 20, no. 4, pp. 339–341, 2000

    2000

  79. [87]

    Urban parcel grouping method based on urban form and functional connectivity characterisation,

    P. Wu, S. Zhang, H. Li, P. Dale, X. Ding, and Y. Lu, “Urban parcel grouping method based on urban form and functional connectivity characterisation,” ISPRS Int. J. Geo-Information, vol. 8, no. 6, p. 282, 2019

    2019

  80. [88]

    MNCD-KE: a novel framework for simultaneous attribute- and interaction-based geographical regionalization,

    L. Xu et al., “MNCD-KE: a novel framework for simultaneous attribute- and interaction-based geographical regionalization,” Int. J. Geogr. Inf. Sci., vol. 38, no. 110, pp. 2148– 2182, 2024. 26

    2024

  81. [89]

    The morphology of urban land use,

    M. Batty and P. A. Longley, “The morphology of urban land use,” Environ. Plan. B Urban Anal. City Sci., vol. 15, no. 4, pp. 461–488, Nov. 1988

    1988

  82. [90]

    Inequality is rising where social network segregation interacts with urban topology,

    G. Tóth et al., “Inequality is rising where social network segregation interacts with urban topology,” Nat. Commun., vol. 12, no. 1, p. 1143, Sep. 2021

    2021

  83. [91]

    An approach for the classification of urban building structures based on discriminant analysis techniques,

    S. Steiniger, T. Lange, D. Burghardt, and R. Weibel, “An approach for the classification of urban building structures based on discriminant analysis techniques,” Trans. GIS, vol. 12, no. 1, pp. 31–59, 2008

    2008

  84. [92]

    Structure and information in spatial segregation,

    P. S. Chodrow, “Structure and information in spatial segregation,” Proc. Natl. Acad. Sci. U. S. A., vol. 114, no. 44, pp. 11591–11596, 2017

    2017

  85. [93]

    A framework for mixed- use decomposition based on temporal activity signatures extracted from big geo-data,

    L. Wu, X. Cheng, C. Kang, D. Zhu, Z. Huang, and Y. Liu, “A framework for mixed- use decomposition based on temporal activity signatures extracted from big geo-data,” Int. J. Digit. Earth, vol. 13, no. 6, pp. 708–726, 2020

    2020

  86. [94]

    A Theory of the Urban Land Market,

    W. Alonso, “A Theory of the Urban Land Market,” Pap. Reg. Sci., vol. 6, no. 1, pp. 149–157, Jan. 1960

    1960

  87. [95]

    Multiple equilibria and structural transition of non-monocentric urban configurations,

    M. Fujita and H. Ogawa, “Multiple equilibria and structural transition of non-monocentric urban configurations,” Reg. Sci. Urban Econ., vol. 12, no. 2, pp. 161–196, May 1982

    1982

  88. [96]

    Krugman, The self organizing economy

    P. Krugman, The self organizing economy. John Wiley & Sons, 1996

    1996

  89. [97]

    The max-p-compact-regions problem,

    X. Feng, S. Rey, and R. Wei, “The max-p-compact-regions problem,” Trans. GIS, vol. 26, no. 2, pp. 717–734, 2021

    2021

  90. [98]

    Fundamental trade-offs on green wireless networks,

    Y. Chen, S. Zhang, S. Xu, and G. Y. Li, “Fundamental trade-offs on green wireless networks,” IEEE Commun. Mag., vol. 49, no. 6, pp. 30–37, Jun. 2011

    2011

  91. [99]

    Study of different resource allocation scheduling policy in advanced LTE with carrier aggregation,

    R. Gatti and S. Hivashankar, “Study of different resource allocation scheduling policy in advanced LTE with carrier aggregation,” in 2017 2nd IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology, 2017, pp. 2257–2261

    2017

  92. [100]

    The delimitation of functional regions, nodal regions and hierarchies by functional distance approaches,

    L. A. Brown and J. Holmes, “The delimitation of functional regions, nodal regions and hierarchies by functional distance approaches,” J. Reg. Sci., vol. 11, no. 1, pp. 57–72, 1971

    1971

  93. [101]

    Network-based functional regions,

    C. J. Q. Farmer and A. S. Fotheringham, “Network-based functional regions,” Environ. Plan. A, vol. 43, no. 11, pp. 2723–2741, 2011

    2011

  94. [102]

    MSDCW: A Dataset of Multi-level Street-block Divisions of 985 Cities Worldwide (Under Review),

    J. Tang et al., “MSDCW: A Dataset of Multi-level Street-block Divisions of 985 Cities Worldwide (Under Review),” Sci. Data, 2025

    2025

  95. [103]

    Inference and analysis across spatial supports in the big data era: Uncertain point observations and geographic contexts,

    C. Robertson and R. Feick, “Inference and analysis across spatial supports in the big data era: Uncertain point observations and geographic contexts,” Trans. GIS, vol. 22, no. 2, pp. 455–476, Apr. 2018

    2018

  96. [104]

    Plot-based urbanism: Towards time-consciousness in place-making,

    S. Porta and O. Romice, “Plot-based urbanism: Towards time-consciousness in place-making,” Glasgow, 2010

    2010

  97. [105]

    Quantitative analysis of urban form: a multidisciplinary review,

    K. Clifton, R. Ewing, G. J. Knaap, and Y. Song, “Quantitative analysis of urban form: a multidisciplinary review,” J. Urban. Int. Res. Placemaking Urban Sustain. , vol. 1, no. 1, pp. 17–45, 2008

    2008

  98. [106]

    Notes on continuous stochastic phenomena,

    P. A. P. Moran, “Notes on continuous stochastic phenomena,” Biometrika , vol. 37, no. 1/2, pp. 17–23, Jun. 1950

    1950

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.