{"id":"92bcfb7f-e4d6-43d5-ad51-2f566fdbf7cd","arxiv_id":"2606.20163","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":4.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Shared EV fleets achieve demand charge savings that recover ownership costs when modeled with spatio-temporal constraints and a marginal-value heuristic.","lead":"This paper models shared EV fleets as mobile batteries to cut demand charges on commercial electricity bills, using a MILP that includes driver labor, transit energy, and battery wear. A generalist might read it to gauge whether real-world logistics make such fleets profitable versus stationary storage.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.3","headline":"Profitability hinges on SF-specific tariffs/demand data; no cross-city validation shown for the economic recovery claim","rationale":"Reader's weakest assumption directly identifies the data/tariff representativeness issue as load-bearing for the profitability conclusion. Full text does not remove this; it supplies the SF-specific derivation but leaves the same external-validity gap. No other internal inconsistency (e.g., in the MILP formulation or heuristic gap) appears more central to the economic claim.","tokens_in":1641,"tokens_out":308,"duration_ms":23006,"concrete_test":"Re-run the MILP and heuristic on an independent dataset (e.g., New York or Los Angeles commercial load profiles + local utility tariffs) using identical cost parameters; if net savings fail to recover TCO for fleet sizes that worked in SF, the headline economic result is location-specific.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the modeled demand-charge savings exceed ownership + operational costs (labor, degradation) for modest fleet sizes. This is computed via MILP/heuristic on San Francisco real-world data and tariffs. The paper examines influence of tariff structures and fleet size within that dataset, but the load-bearing condition is that outcomes are not artifacts of SF load shapes, peak timing, or rate design. If the same framework on other cities' data yields negative net present value, the claim that such fleets are techno-economically viable does not hold beyond the single location.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper develops a high-fidelity MILP formulation for shared EV fleet dispatch that jointly minimizes demand charges and total cost of ownership while incorporating spatio-temporal energy use, driver labor costs, and battery degradation. A marginal-value heuristic is proposed to solve the path-dependent problem efficiently. Using San Francisco real-world load and tariff data, the analysis concludes that modest fleet sizes can generate demand-charge savings sufficient to recover ownership and operating costs, with further results on sensitivity to tariff design and fleet size.","tokens_in":1740,"tokens_out":394,"duration_ms":11350,"significance":"If the numerical results and economic recovery claim hold under scrutiny, the work supplies a practical, constraint-aware framework that moves beyond idealized mobile-storage models; the explicit inclusion of labor and degradation costs and the near-optimal heuristic are concrete strengths that could inform fleet operators and tariff design.","major_comments":[{"comment":"The central techno-economic claim (abstract and results) that savings recover ownership plus operational costs rests entirely on San Francisco load shapes, peak timing, and rate structures. No cross-city validation or alternative datasets are presented; if the same MILP/heuristic on other urban data yields negative NPV, the viability conclusion does not generalize beyond the single location.","section":"Abstract and Results section"},{"comment":"The profitability statements depend on fitted cost parameters (labor, degradation) and the specific tariff; the manuscript does not provide external benchmarks or sensitivity ranges that would allow readers to assess whether the net-positive outcome is robust or an artifact of the chosen data source.","section":"Results section"}],"minor_comments":[{"comment":"Notation for the marginal-value heuristic and the MILP objective could be clarified with an explicit list of decision variables and constraints in one location.","section":"Methods"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the detailed and constructive review. The comments correctly identify that our analysis is a single-city case study and that parameter robustness merits further attention. We respond to each major comment below and indicate the revisions we will make.","responses":[{"response":"We agree that the numerical results and economic-recovery conclusion are tied to San Francisco load shapes, tariffs, and operating conditions; the manuscript presents a detailed case study rather than a multi-city generalization. The MILP formulation and marginal-value heuristic are data-agnostic and can be re-parameterized for other cities, but we do not possess additional urban datasets at this time. We will revise the abstract, introduction, and conclusions to explicitly frame the work as a San Francisco case study, add a dedicated limitations subsection discussing transferability, and include qualitative discussion of how load-profile and tariff differences in other cities would affect outcomes.","revision_made":"partial","referee_comment":"[Abstract and Results section] The central techno-economic claim (abstract and results) that savings recover ownership plus operational costs rests entirely on San Francisco load shapes, peak timing, and rate structures. No cross-city validation or alternative datasets are presented; if the same MILP/heuristic on other urban data yields negative NPV, the viability conclusion does not generalize beyond the single location."},{"response":"The current manuscript already reports sensitivity to tariff design, fleet size, and selected cost components. To strengthen the robustness assessment we will (i) expand the sensitivity ranges for labor and degradation parameters using values drawn from the cited literature, (ii) add a table comparing our base-case parameters against external benchmarks from fleet-operator reports and prior EV studies, and (iii) include additional tornado plots showing NPV sensitivity to the most uncertain inputs.","revision_made":"yes","referee_comment":"[Results section] The profitability statements depend on fitted cost parameters (labor, degradation) and the specific tariff; the manuscript does not provide external benchmarks or sensitivity ranges that would allow readers to assess whether the net-positive outcome is robust or an artifact of the chosen data source."}],"tokens_in":1258,"tokens_out":451,"duration_ms":24276,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"This paper applies an existing MILP formulation for EV dispatch to a shared fleet setting and adds explicit terms for driver labor and battery degradation. With San Francisco real-world data and tariffs, it finds that a small number of vehicles can produce demand charge savings large enough to cover ownership and operating costs. The main addition is the high-fidelity treatment of those practical costs plus a marginal-value heuristic that solves the problem faster while staying close to optimal.\n\nThe modeling choices are reasonable. Including transit overhead, labor, and degradation makes the setup more usable than idealized versions that ignore them. The sensitivity results on fleet size, tariff design, and cost components are straightforward and give operators a usable picture of what moves the economics.\n\nThe limitation is the single-city scope. Profitability rests on San Francisco load shapes and rate structures; the paper varies parameters inside that dataset but does not run the same framework on data from other cities. If peaks or tariffs differ elsewhere, the net savings could disappear, so the viability claim does not travel far without further checks. No external benchmarks or parameter-free derivations appear.\n\nThis is the sort of applied case study that facility managers or EV fleet operators might use when planning demand response in similar settings. Readers wanting new optimization theory or broadly validated results will not find them here.\n\nI would not cite it unless the application matches the SF conditions closely. It should go to peer review so referees can examine the MILP details, heuristic performance, and data handling.","headline":"The paper runs a MILP on SF data to show modest shared EV fleets can recover costs via demand charge cuts once labor and degradation are included, but the economics stay tied to that location.","tokens_in":2233,"tokens_out":382,"would_cite":false,"duration_ms":29679,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Shared electric vehicle fleets can reduce demand charges enough to recover their full ownership and operational costs.","keywords":["electric vehicles","demand charge reduction","shared mobile storage","techno-economic analysis","mixed-integer linear programming","fleet dispatch","battery degradation"],"falsifier":"Re-running the MILP and heuristic on electricity consumption and tariff data from a different city and finding that the net savings fall short of recovering ownership and operational costs.","tokens_in":2527,"feed_emoji":"🚗","tokens_out":581,"duration_ms":15637,"temperature":0.7,"pith_summary":"This paper examines whether fleets of electric vehicles can be shared as mobile storage to lower demand charges on electricity bills. It builds a detailed model that includes the time and cost of moving the vehicles between locations, driver wages, and how much the batteries wear out. The analysis uses real data from San Francisco to show that even a small number of these vehicles can generate savings large enough to cover all expenses. This matters because it suggests a practical way for EV owners to earn money from grid services without ignoring real-world logistics.","feed_headline":"Shared EV fleets recover costs by cutting demand charges","feed_subtitle":"San Francisco data shows modest fleets offset ownership and labor expenses when transit and degradation are modeled explicitly.","key_machinery":"The mixed-integer linear program (MILP) that jointly minimizes demand charges and total cost of ownership, supported by a marginal-value-based heuristic algorithm for efficient solving of the dispatch problem.","core_discovery":"The central claim is that a modest number of shared EVs, managed with a high-fidelity framework accounting for spatio-temporal energy use, labor costs, and battery degradation, can achieve demand charge savings sufficient to recover ownership and operational expenses when applied to San Francisco real-world data and tariff structures.","pith_inferences":["If the cost-recovery result generalizes, shared EV fleets could serve as a bridge technology that improves the economics of vehicle-to-grid participation.","Similar modeling could be extended to include interactions with on-site solar or wind generation to increase total value.","Sensitivity tests across battery chemistries with different degradation curves would clarify the robustness of the profitability threshold."],"forward_implications":["Tariff structures directly influence the overall profitability of the shared EV fleet.","Fleet size determines the scale of achievable demand charge savings and net returns.","Labor costs for drivers and battery degradation rates affect whether the operation breaks even.","The heuristic algorithm enables practical, near-optimal dispatch decisions at scale."],"fun_headline_variants":["EV fleets cut demand charges to recover ownership costs","Shared EVs in San Francisco offset costs with demand savings","Modest EV fleets recover costs via demand charge cuts","Demand charge cuts cover EV fleet expenses according to SF data"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The assumption that San Francisco electricity tariffs and demand patterns are representative of conditions in other locations where such shared EV fleets might operate.","fun_headline_variants_meta":{"raw":{"variants":["EV fleets cut demand charges to recover ownership costs","Shared EVs in San Francisco offset costs with demand savings","Modest EV fleets recover costs via demand charge cuts","Demand charge cuts cover EV fleet expenses according to SF data"]},"model":"grok-4.3","cost_usd":0.008564,"raw_usage":{"total_tokens":3819,"prompt_tokens":572,"num_sources_used":0,"completion_tokens":61,"cost_in_usd_ticks":85637000,"prompt_tokens_details":{"text_tokens":572,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":3186,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":572,"tokens_out":61,"duration_ms":29685,"temperature":1.0,"reasoning_tokens":3186,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-26T16:15:22.540059+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Re-running the MILP and heuristic on electricity consumption and tariff data from a different city and finding that the net savings fall short of recovering ownership and operational costs.","supporting_citations":[],"review_version":1}