{"submission_id":"b449c17e-8b4c-44d5-99d4-8c092a906d5b","submission_revision_id":"1a106017-695d-4ea8-8765-7300175f2667","revision_version":1,"submission_status":"under_review","submitted_at":"2026-09-07T14:04:18.395107Z","submitter_name":"Yufan Zhang","submitter_email":"yufanz@umd.edu","task":{"status":"submitted","version":1,"extra_fields":{},"submitted_at":"2026-09-07T14:04:18.395107+00:00","industry_domain":"Transportation","task_short_name":"sumo_special_event_egress_1","operating_system":"Linux, MacOS","task_description":"Objective:\nDesign a robust post-event traffic-management plan for clearing 25,000 spectators from six parking lots through a signalized urban road network. Minimize network clearance time, event-vehicle travel time, and maximum queue length.\n\nStarting State:\nThe input package contains five immutable Eclipse SUMO 1.26.0 network profiles representing a 4-by-4 urban grid with 16 fixed-time traffic signals, six parking lots with two physical gates each, four regional gateways, 6,250 event vehicles representing 25,000 spectators, and 690 background vehicles in the visible scenario. It also includes the parking-lot and gateway contract, the visible demand data, a valid starter solution, and the required JSON schema.\n\nInstructions & Rules:\nModify only public/solution.json.\n\nFor each parking lot, specify:\n- the release start delay;\n- the metered vehicle-release rate;\n- the proportion assigned to gate A versus gate B;\n- the proportion assigned to the lot’s primary versus secondary regional gateway.\n\nAlso specify:\n- the proportion of event vehicles assigned to an alternate route;\n- whether event vehicles are restricted from making left turns at the four central intersections;\n- whether one north, east, south, or west corridor is temporarily converted to outbound one-way operation.\n\nAll values must satisfy the ranges and types defined in public/solution.schema.json. Do not modify the road networks, traffic signals, vehicle demand, parking-lot occupancy, vehicle IDs, background traffic, gateway definitions, simulation duration, hidden scenarios, or evaluator. Vehicles may not be removed or manipulated to improve the result.\n\nEvery evaluated scenario must deliver all event and background vehicles, produce zero teleports, clear event traffic within 9,850 seconds, and keep the maximum lane queue at or below 280 meters.\n\nFinal evaluation includes the visible scenario and three private scenarios containing undisclosed variations in total attendance, parking-lot occupancy distribution, and background traffic demand.\n\nDeliverable:\nSubmit the completed public/solution.json file. The private evaluator will deterministically compile the submitted plan into vehicle departures and routes, rerun SUMO, and assign a continuous score from 0 to 100.","software_licensing":"Free/Open Source","evaluation_criteria":null,"verification_method":"Evaluate the submission by deterministic SUMO performance rather than exact equality with the reference JSON.\n\n1. Submission validation\n\nParse public/solution.json and require it to match the supplied schema exactly. Missing or additional fields are invalid.\n\nFor each of L0 through L5, require:\n- start_delay_s: integer from 0 to 900\n- release_rate_veh_per_min: integer from 8 to 30\n- gate_a_share: finite number from 0 to 1\n- primary_gateway_share: finite number from 0 to 1\n\nRequire:\n- alternate_route_share: finite number from 0 to 0.5\n- restrict_event_left_turns: Boolean\n- temporary_one_way: one of none, north, east, south, or west\n\nConfirm that all fixed networks, demand files, parking-lot definitions, traffic signals, vehicle records, scenario data, and evaluation files remain unchanged.\n\n2. Deterministic simulation\n\nCompile the submitted release schedule, physical-gate allocation, regional-gateway allocation, route distribution, movement restriction, and temporary one-way selection deterministically from the supplied vehicle IDs.\n\nRun Eclipse SUMO 1.26.0 on:\n- the visible scenario with weight 0.40;\n- three private scenarios with weight 0.20 each.\n\nIdentical inputs and submissions must produce identical results.\n\n3. Scenario metrics and cost\n\nFor every scenario, record:\n- network_clearance_time_s\n- event_mean_travel_time_s\n- max_queue_m\n- event_vehicles_arrived\n- background_vehicles_arrived\n- vehicles_arrived\n- teleports\n\nCalculate:\n\nscenario_cost = network_clearance_time_s\n              + 0.50 * event_mean_travel_time_s\n              + 0.20 * max_queue_m\n\nLower cost is better.\n\n4. Hard constraints\n\nEvery individual scenario must satisfy:\n- event_vehicles_arrived equals the expected event-vehicle count\n- background_vehicles_arrived equals the expected background-vehicle count\n- teleports equals 0\n- network_clearance_time_s is no more than 9,850 seconds\n- max_queue_m is no more than 280 meters\n\nIf any scenario violates any hard constraint, the final score is 0.\n\n5. Weighted evaluation\n\nCalculate:\n\nweighted_cost = 0.40 * visible_cost\n              + 0.20 * hidden_cost_1\n              + 0.20 * hidden_cost_2\n              + 0.20 * hidden_cost_3\n\n6. Continuous score\n\nFor a valid and feasible submission, calculate:\n\nscore = 100 * (9017.074600 - weighted_cost)\n              / (9017.074600 - 7589.246300)\n\nRound the score to three decimal places and clip it to the range [0, 100].\n\nThe conservative starter plan anchors a score of 0. The fixed feasible expert reference, with a weighted cost of 7589.246300, anchors a score of 100. Intermediate feasible performance receives a proportional intermediate score.\n\nInvalid JSON, modified fixed inputs, failed simulation, missing or non-finite metrics, or any hard-constraint violation receives 0. If multiple submissions reach the 100-point cap, use the lower raw weighted cost as the tie-breaker.","software_and_version":"Eclipse SUMO 1.26.0, Python 3.11, sumolib 1.26.0, Pixi","specific_requirements":null,"expected_completion_time":null},"files":[{"file_id":"87a3f805-0047-45e1-80d3-bb0539f9b6d4","filename":"sumo_special_event_egress_1_input.zip","category":"input_materials","content_type":"application/zip","size_bytes":119189},{"file_id":"8d768cd2-6f50-47c1-9567-fafef4f12125","filename":"sumo_special_event_egress_1_reference.zip","category":"reference_output","content_type":"application/zip","size_bytes":141253}],"deep_review":{"overall_verdict":"Strong Accept","review_text":"**Overall Verdict:** Strong Accept — The submission is complete, coherent, objectively evaluable, and reachable through a deterministic, professionally representative SUMO workflow.\n\n**Completeness Check:**\n- Instruction: Complete — The decisions, permitted ranges, immutable inputs, hard constraints, scoring objective, and required deliverable are explicit.\n- Inputs: Complete — Five network profiles, visible demand and metadata, six-lot gateway contract, starter solution, and exact JSON schema are present and internally consistent.\n- Toolchain: Complete — Python 3.11 and SUMO 1.26.0 dependencies are pinned with a Linux lockfile, environment checks, and evaluation entry point.\n- Reference output: Complete — The expert solution, baseline and expert four-scenario reports, private scenarios, and verification guide are supplied separately from solver-visible inputs.\n- Rubric: Complete — Validation, scenario weights, metrics, hard-failure conditions, cost formula, score normalization, rounding, clipping, and tie-breaking are fully specified and implemented consistently.\n\n**File and Task Reachability:**\nAttachment classification: `reachable`. Task classification: `reachable`. Both `sumo_special_event_egress_1_input.zip` and `sumo_special_event_egress_1_reference.zip` were inspected. They are valid, unencrypted ZIP archives matching their declared sizes, with 16 and 46 entries respectively, no unsafe paths, links, devices, or excessive expansion. Every member was inventoried; the JSON, Markdown, Python, shell, TOML, lockfile, and representative XML content were inspected. All JSON parsed, declared artifact hashes matched, solver-visible files agreed across packages, both solutions satisfied the schema contract, and the published scenario costs and weighted anchors recomputed exactly. The five network files contain the declared 16 signal programs, required lot and gateway edges, and directional four-lane conversions with reverse-edge closure. No material content was omitted. A live SUMO rerun and independent XML-schema validation were unavailable, so frozen simulation metrics were verified structurally and arithmetically rather than reproduced; this is a reviewer-tooling limitation, not a submission defect. End to end, the contributor edits only `public/solution.json`; the evaluator validates it, deterministically compiles departures and routes, runs four scenarios, checks arrivals, teleports, clearance and queues, then computes the weighted score. License keys, accounts, tokens, and other credentials are not required from the contributor. Confidential, proprietary, or internal material may be evaluated as submitted evidence without requiring additional disclosure.\n\n**Special Screening:**\nNone — no affirmative evidence supports a screening label; public provenance could not be independently established from the submitted evidence.\n\n**Recommended Changes:**\nThe submission provides a well-separated public task and private evaluation package with reproducible scoring evidence. No changes are required.","reviewed_at":"2026-09-07T14:16:40.069638Z"},"claim":null}