{"id":"82aed6fb-6b69-4117-8816-63d2f83f8c11","arxiv_id":"2501.17873","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A quality-of-service resource allocation framework with a new adaptive fast traversal algorithm and 3D strip packing lets a split-aperture radar track more targets simultaneously than a conventional full-aperture scheduler.","lead":"This paper presents a radar resource management scheme for split-aperture phased arrays, where the antenna is divided into subarrays that run tracking tasks simultaneously. The authors show in simulation that this can support more active tracks than scheduling tasks on the full aperture one at a time.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Simulation benefit hinges on an admittedly arbitrary cross-talk model (Eq. 5); its ~1 dB worst-case penalty is an untested assumption that directly sets the split-aperture advantage.","rationale":"I read the paper as a coherent engineering contribution: it formulates a Q-RAM/KBS-based resource allocation problem for split-aperture arrays, introduces a 3D strip-packing model for the coupled scheduling constraint, and proposes an adaptive fast traversal algorithm to handle the resulting discrete jumps. The simulation results are internally consistent in that the constrained split-aperture curve lies between the unconstrained split-aperture curve and the full-aperture baseline. However, the central claim of significantly increased active tracks depends on the cross-talk loss model, and the paper itself states that this model is chosen arbitrarily because the front-end cross-talk problem is open. This is not an internal inconsistency, but it is an unsupported modeling assumption directly in the path from the SAPA concept to the simulated benefit. The cross-talk penalty affects both detection/quality and the expected number of looks, so even a few dB of additional loss could shift the resource requirements enough to close the gap shown in Fig. 5. The heuristic packing algorithm and AFT hyperparameters are secondary concerns: they affect how close the constrained result is to the unconstrained ideal, but they would not overturn a fundamental split-versus-full comparison if the physical loss model were sound. The absence of code and data is a reproducibility issue, not the decisive scientific concern. Since the reader already reached CONDITIONAL and the key vulnerability is the same one the reader identified, I would not change the verdict. The paper should either ground Eq. (5) in measurements or a more physical model, or explicitly present the conclusion as conditional on that arbitrarily chosen model.","tokens_in":15727,"tokens_out":6984,"duration_ms":78816,"concrete_test":"Rerun the Section VI Monte Carlo with xi = c + (1-c)*(Nh,k/NhT)*(Nv,k/NvT) for c swept over, say, [0, 0.95], and report the active-track curves. Identify the threshold c* at which the constrained split-aperture mean active-track count no longer exceeds the full-aperture baseline by the margin stated in the conclusion. If c* is below a value consistent with measured subarray isolation, or if a full-wave EM simulation of the 48x48 array with simultaneous subarray beams gives a larger loss, the conclusion should be reworded as conditional on the cross-talk assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is a simulation demonstration, and every curve in Figs. 5-7 depends on Eq. (5), the cross-talk loss. The paper explicitly declares this function arbitrary (footnote 1). It sets xi = 0.8 for a vanishingly small subarray and xi = 1 for the full array, i.e., at most about 1 dB penalty for splitting. Because xi enters beta = xi*SN0 - ln(P_fa) in Eq. (4b) and also P_D and n_l in Eqs. (9b)-(9c), it controls both the quality achievable with a given aperture and the resource a task consumes. A larger or physically different cross-talk penalty would increase the resource needed by small-aperture tasks and could reduce or eliminate the active-track advantage over the full-aperture baseline. The paper provides no measurement, EM simulation, or citation to support the 0.8 floor, so the claimed 'significant increase' is not yet robust to the one model ingredient that is unique to SAPA.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a radar resource management formulation for split-aperture phased array (SAPA) systems. Tracking task quality is modeled with the Van Keuk and Blackman strategy, task resource consumption includes coherent integration time, update rate, and aperture size, and concurrent tasks are scheduled over the array by solving a three-dimensional strip packing problem. The resulting quality-of-service allocation problem is solved with a new adaptive fast traversal algorithm combined with a 3SP heuristic. A Monte Carlo simulation with 60 targets and two range configurations compares unconstrained split-aperture, constrained split-aperture, and full-aperture sequential scheduling, and the authors conclude that the SAPA concept can significantly increase the number of active tracks for the same radar time budget.","tokens_in":15958,"tokens_out":5315,"duration_ms":61724,"significance":"If the demonstrated benefit is robust, the paper makes a useful contribution: it is, to the authors' knowledge, the first work to include constrained array scheduling in SAPA resource management, and it connects the problem to the Q-RAM framework and the 3SP literature. The simulation study is transparent about its settings, uses external models (KBS and Q-RAM) rather than constants fitted to the reported outcome, and shows consistent trends across Monte Carlo runs. The main weakness is that the central quantitative claim rests on an admittedly arbitrary cross-talk loss model and on two heuristic solution components whose approximation quality is not characterized in the paper.","major_comments":[{"comment":"The cross-talk loss function ξ = 0.8 + 0.2 (Nh,k/NhT)(Nv,k/NvT) is explicitly declared 'chosen arbitrarily' in footnote 1. This function enters β in Eq. (4b), and through PD and nl in Eqs. (9b)-(9c) it affects both the quality qk and the resource gk of every split-aperture task. The assumed model penalizes the smallest subarrays by only about 1 dB relative to the full aperture, and this penalty directly controls the size of the split-aperture advantage reported in Figures 5-7. A larger or physically different cross-talk penalty would increase the resources required by small-aperture tasks and could reduce or eliminate the active-track benefit over the full-aperture baseline. The paper needs either a physical or measured justification for this loss model or a sensitivity analysis over its parameters before the central claim can be considered supported.","section":"Eq. (5) and footnote 1"},{"comment":"The adaptive fast traversal algorithm has four user-chosen parameters α1, n1, n2, and n3, which are set to 0.7, 2, 3, and 3 in Section VI.A without any sensitivity analysis; Section VII lists studying the impact of these parameters as future work. Because Algorithm 3 only approximates the concave majorant and the marginal-utility ordering determines which tasks receive resources, the reported active-track, utility, and computation-time results are conditional on an uncharacterized approximation accuracy. A parameter-sensitivity study, or a comparison against full enumeration for a reduced control space, is needed to show that the reported benefits are not artifacts of this particular choice of AFT hyperparameters.","section":"Algorithm 3 and Section VI.A"},{"comment":"The split-aperture resource function g(S*) is the height produced by the DBLF-plus-shaking heuristic of [24], and no optimality bound is provided for this heuristic on the task sets used in the simulation. Since this resource value enters the marginal utility computation in Algorithm 3, suboptimal packings can affect not only the final resource consumption but also which allocations are selected by the resource management algorithm. The paper should quantify the heuristic's quality on small instances against an exact 3SP solver or on standard strip-packing benchmarks, so the constrained curves in Figures 5-7 can be interpreted with known confidence.","section":"Section IV.A and Eq. (11)"}],"minor_comments":[{"comment":"The parameters for targets T1 and T2 appear inconsistent with Table I: the text assigns Σk = 35 m/s2 and Θk = 10 s for T1, but Table I lists Θ as the standard deviation in m/s2 and Σ as the time correlation in s. Please check the notation and the numeric assignments.","section":"Section VI.A and Table I"},{"comment":"Footnote 4 is truncated and the surrounding sentence is garbled ('target T1 only exists for short ranges as a target cannot be .'), which makes the intended explanation unclear.","section":"Section VI.A, footnote 4"},{"comment":"The use of uk and gk in Eq. (14) is confusing because S is a set of index points for all tasks; the marginal utility should be defined using the total utility u and the coupled resource function G, or the notation should be clarified explicitly.","section":"Eq. (14)"},{"comment":"In the paragraph describing the unconstrained case, 'without contains' should read 'without constraints'.","section":"Section VI.A"},{"comment":"The algorithm named 'CPADS' in the final paragraph is elsewhere called CDAPS (reference [28]); please make the acronym consistent.","section":"Section VI.B"},{"comment":"There are several minor language issues, including 'an 3 dimensional strip packing', 'the concave-majorant', and 'the resource allocation problem in (1) is solved using the following steps'; a careful copyedit would improve readability.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is already accepted at IEEE TAES, and this report is therefore a post-acceptance assessment. The novelty claim is not in question; the main issue is that the quantitative central claim depends on an admittedly arbitrary cross-talk model and on heuristic components whose approximation error is not bounded or studied. These are fixable within the manuscript's scope through sensitivity analysis and a benchmark-based validation of the 3SP and AFT components, which is why I recommend major revision rather than rejection. If the venue prefers post-publication commentary, the major comments could instead be addressed in a correction or follow-up paper."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Friendly take: this is a solid engineering contribution, and I'd send it to review. The genuinely new pieces are the coupled split-aperture scheduling formulation as a 3D strip-packing problem and the adaptive fast traversal algorithm that handles discrete jumps in the marginal-utility approximation. The combination with Q-RAM/KBS is sensible, and the simulation study is transparent: 100 Monte Carlo runs, two target-range scenes, and comparisons against both an unconstrained split-aperture bound and the full-aperture baseline. The authors are also honest about the computational overhead and about the places where their model is not grounded.\n\nThe main soft spot, which the stress-test note gets right, is the cross-talk loss function in Eq. (5). The paper itself says it is 'chosen arbitrarily' because modeling front-end cross-talk in SAPA is open. That function directly controls SNR via Eq. (4b) and consequently the resource function, so the size of the claimed active-track advantage is not robust until someone checks sensitivity to that assumption. A simple parameter sweep varying the floor (0.8) and the shape of xi would go a long way. The 3SP solver is a heuristic with no optimality bound, and the AFT hyperparameters (alpha1, n1, n2, n3) are set without sensitivity analysis; the authors list that as future work. These are real limitations but they weaken the quantitative claims, not the formulation.\n\nI do not think the central framework is circular: the baselines come from prior external models and the 3SP heuristic from the OR literature. The paper is careful to label what is new and what is assumed.\n\nWho should read it: radar resource management researchers and engineers working on digital array front-ends. It is a niche contribution but a meaningful one. It deserves a serious referee; the referee should ask for a sensitivity analysis on the cross-talk model and AFT parameters, and ideally for code/data to reproduce the simulation. I would not desk-reject it.","headline":"Real contribution to radar resource management, but the headline track-count gain hinges on an admittedly arbitrary cross-talk loss model.","tokens_in":16445,"tokens_out":2958,"would_cite":true,"duration_ms":30076,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A radar that divides its array into sub-arrays, each running a different tracking task at the same time, can keep more target tracks alive for the same radar time budget than a radar that schedules the same tasks sequentially on the full…","keywords":["Cognitive radar","Radar resource management","Radar tracking","Split-aperture phased arrays","Task dependencies","Q-RAM","Adaptive fast traversal algorithm","Three-dimensional strip packing"],"falsifier":"Replace the arbitrary cross-talk model in equation (5) with a measured or electromagnetically simulated inter-subarray coupling loss and re-run the 70 km Monte Carlo scenario; if the constrained split-aperture active-track curve falls to or below the full-aperture curve anywhere in the 0\\textendash20% radar time budget range, the paper's central claim is refuted.","tokens_in":15531,"feed_emoji":"📡","tokens_out":14679,"duration_ms":135879,"temperature":0.7,"pith_summary":"This paper argues that a phased-array radar whose aperture can be split into sub-arrays, each executing a different tracking task at the same time, can keep significantly more active target tracks alive for the same radar time budget than a radar that schedules those tasks one after another on the full aperture. The authors formulate split-aperture resource management as a quality-of-service optimization, pairing an empirically calibrated tracking model that sets track quality as angular estimation error with a three-dimensional strip packing of tasks into the array's element-and-time volume. Because packed tasks share the array, the resource function becomes coupled and the utility-versus-resource curve develops discrete jumps, so the paper introduces an adaptive fast traversal algorithm to solve the allocation despite those jumps. A Monte Carlo simulation with 60 targets shows the constrained split-aperture solution keeping more tracks active at every tested time budget than the full-aperture baseline, with its performance between the unconstrained split-aperture ideal and the sequential baseline. If the result survives contact with hardware, future multifunction radars could serve more simultaneous missions without increasing transmit power or total radar time.","feed_headline":"Splitting a radar array raises its active track count","feed_subtitle":"Simulation with 60 targets shows the split array keeps more tracks alive at every radar time budget tested.","key_machinery":"The mechanism has three parts that work together. First, the quality-of-service based resource allocation model (Q-RAM) formulates the problem as maximizing weighted task utility under a total radar time budget, with each task's utility derived from the empirically calibrated tracking strategy (KBS) that sets quality equal to the angular estimation error of the track. Second, each task is a box with dimensions (horizontal elements $N_{h,k}$, vertical elements $N_{v,k}$, expected radar time $g_k$) that must be packed without overlap into the array's fixed element area, making the coupled resource function $g(S^\\star) = \\max_{k \\in K} (z_k + g_k)$, i.e., the highest top of any box in the stack; packing is done by the deepest-bottom-left-fill heuristic with a shaking procedure. Third, because packing couples tasks, the utility-versus-resource curve has discrete jumps that the standard fast traversal method cannot handle, so the paper introduces the adaptive fast traversal algorithm, which searches tree branches around first-order candidate points for equal-resource, higher-utility alternatives and prunes that search adaptively via the parameters $\\alpha_1, n_1, n_2, n_3$.","core_discovery":"The paper's central claim, stated in its own conclusion, is that the split-aperture phased array concept can significantly increase the number of active tracks of a multifunction radar system compared to scheduling tasks sequentially. The demonstration is a 100-run Monte Carlo simulation of 60 tracking tasks in two scenes (targets out to 70 km and 250 km), comparing three allocation problems: split-aperture without packing constraints, split-aperture with the three-dimensional strip packing constraints, and full-aperture sequential allocation. On the paper's own terms, the constrained split-aperture solution keeps more tracks alive at every radar time budget tested, with total utility and mean angular estimation error ordered the same way, and the authors further state this is the first treatment of constrained scheduling over the array for split-aperture resource management. The claim is that the extra flexibility of dividing the array into sub-arrays, giving short-range or low-maneuver targets only the elements they need, translates directly into more simultaneous tracks for the same time budget.","pith_inferences":["If the arbitrary cross-talk model in equation (5) is replaced by a physically computed inter-subarray coupling, the reported gain will change; the current simulation conflates real cross-talk losses with the packing-induced losses that the strip packing algorithm is meant to measure.","The large gap in computation time between the unconstrained and constrained solvers suggests an anytime design: serve the unconstrained allocation immediately and refine it with packing constraints within the revisit interval, which the paper does not explicitly propose.","The box-packing view of tasks extends beyond tracking to mixed-function radars, treating search, track, and communication as boxes with different quality functions, so this machinery is a candidate framework for holistic multifunction resource management.","The adaptive fast traversal solver has four user-set parameters ($\\alpha_1, n_1, n_2, n_3$) whose effect on solution quality is left to future work; a sensitivity study would show whether the track-count advantage is robust or tuning-dependent."],"forward_implications":["A multifunction radar with aperture-splitting capability can trade track quality inside its permitted band ($1$\\textendash$3$ mrad angular estimation error) to keep more simultaneous tracks alive, and the simulation shows the mean error stays below the $3$ mrad limit in every configuration.","In the 70 km scene the constrained split-aperture curve nearly reaches the unconstrained ideal, showing that packing losses are small when many tasks need only a small part of the aperture.","The full-aperture allocation is an upper bound on required resource and the unconstrained split-aperture allocation a lower bound, so any physically realizable split-aperture scheduler's track count will lie between these two curves.","Short-range, large-RCS, or slow-maneuvering targets, which need few array elements, are exactly the tasks that free aperture space for other tracks, which is where the track-count gain comes from.","As presented, the constrained split-aperture allocation is not fast enough for real-time operation; the paper lists a faster implementation, a more efficient packing variant, and hot-started CDAPS search as the routes to closing that gap."],"supporting_citations":[{"why":"Supplies the empirically calibrated tracking strategy that defines track quality as angular estimation error and predicts the expected steady-state radar resource demand.","marker":"[15]"},{"why":"Supplies the quality-of-service based resource allocation model (Q-RAM) in which the split-aperture allocation problem is formulated and solved.","marker":"[14]"},{"why":"Supplies the deepest-bottom-left-fill packing heuristic with shaking procedure that places the task boxes on the array.","marker":"[24]"},{"why":"Supplies the fast traversal algorithm that the new adaptive fast traversal extends to handle discrete jumps in marginal utility.","marker":"[27]"},{"why":"Provides the quality and resource equations from multi-mission radar network resource management that this paper adapts to the split-aperture case.","marker":"[19]"},{"why":"The authors' own earlier split-aperture tracking allocation work, which this paper extends by adding task coupling and the adaptive traversal algorithm.","marker":"[30]"},{"why":"Establishes that three-dimensional strip packing is NP-hard in the strong sense, motivating the heuristic packing approach.","marker":"[21]"},{"why":"Supplies the radar equation for signal-to-noise ratio that underlies the quality and resource functions.","marker":"[31]"}],"fun_headline_variants":["A split-aperture radar can track more targets at once","Splitting a radar array into sub-arrays boosts track capacity","Split-aperture radar keeps more tracks alive per time budget","Radar array splitting enables more simultaneous tracks","Split aperture radar achieves higher track capacity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The simulated benefit of splitting the aperture rests on equation (5), the cross-talk loss model $\\xi = 0.8 + 0.2 (N_{h,k}/N_{hT})(N_{v,k}/N_{vT})$, which the paper's own footnote admits was chosen arbitrarily because modeling front-end cross-talk for the split-aperture concept is an open problem.","fun_headline_variants_meta":{"raw":{"variants":["A split-aperture radar can track more targets at once","Splitting a radar array into sub-arrays boosts track capacity","Split-aperture radar keeps more tracks alive per time budget","Radar array splitting enables more simultaneous tracks","Split aperture radar achieves higher track capacity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001127,"raw_usage":{"total_tokens":4666,"prompt_tokens":909,"completion_tokens":3757,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":3679}},"tokens_in":525,"tokens_out":3757,"duration_ms":30446,"temperature":1.0,"reasoning_tokens":3679,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:37:56.590112+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Replace the arbitrary cross-talk model in equation (5) with a measured or electromagnetically simulated inter-subarray coupling loss and re-run the 70 km Monte Carlo scenario; if the constrained split-aperture active-track curve falls to or below the full-aperture curve anywhere in the 0\\textendash20% radar time budget range, the paper's central claim is refuted.","supporting_citations":[{"cited_title":"Van Keuk and S","cited_arxiv_id":null,"evidence_quote":"Supplies the empirically calibrated tracking strategy that defines track quality as angular estimation error and predicts the expected steady-state radar resource demand."},{"cited_title":"Rajkumar, C","cited_arxiv_id":null,"evidence_quote":"Supplies the quality-of-service based resource allocation model (Q-RAM) in which the split-aperture allocation problem is formulated and solved."},{"cited_title":"Wauters, J","cited_arxiv_id":null,"evidence_quote":"Supplies the deepest-bottom-left-fill packing heuristic with shaking procedure that places the task boxes on the array."},{"cited_title":"Ghosh, R","cited_arxiv_id":null,"evidence_quote":"Supplies the fast traversal algorithm that the new adaptive fast traversal extends to handle discrete jumps in marginal utility."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The authors' own earlier split-aperture tracking allocation work, which this paper extends by adding task coupling and the adaptive traversal algorithm."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the radar equation for signal-to-noise ratio that underlies the quality and resource functions."}],"review_version":1}