基于群智演化学习的多目标复杂多智能体博弈模型的新能源汽车共享换电场景博弈数据集内容为小规模（30辆车、5个换电站）、中规模（150辆车、25个换电站）和大规模（600辆车、100个换电站）三类静态完全信息同步博弈场景，
\n每类包含20个独立随机实例，共60个JSON格式实例文件。每个实例记录车辆与换电站的二维坐标（区域边长20）、各站电池数量（小/中规模5~10块，大规模10~15块）、每块电池初始SoC（0.5~1.0均匀采样）、车辆初始SoC（0.2~0.4均匀采样）、车辆最低安全到达SoC（初始SoC的1/3~2/3）、车辆换电需求阈值（0.5~0.7均匀采样）等参数，并基于动态电价模型定义个体成本与社会总成本。
", "ds_source": "本数据集为完全仿真生成数据，不源自特定文献、实测或第三方下载。
", "ds_process_way": "本数据集基于仿真生成方式构建，其生成过程基于Python仿真程序和Numpy随机采样机制。具体而言，首先根据新能源汽车共享换电场景的实验需求，配置不同规模的待换电车辆数量和换电站数量，构造小规模、中规模和大规模三类博弈场景，其中小规模包含30辆EV和5个BSS，中规模包含150辆EV和25个BSS，大规模包含600辆EV和100个BSS，并在每类场景下分别生成20个独立随机实例，共形成60个数据实例。随后，利用随机采样方法在20×20的二维区域内生成EV与BSS的空间坐标，并基于车辆与换电站之间的欧氏距离计算车辆前往不同换电站的行驶距离，为后续能耗、到达安全约束和行驶成本计算提供基础数据。同时，根据不同场景规模设置单个换电站的电池数量上下界，对各换电站可用电池数量进行整数均匀采样；若生成的电池总数少于车辆数量，则重新采样，以保证整体资源规模满足多车辆换电决策实验要求。进一步地，结合单位里程能耗系数、电池额定容量、动态电价参数、换电支付成本权重和行驶成本权重，计算车辆在不同换电选择下的个体成本；对于违反局部可行性约束的选择，在成本函数中加入惩罚项。最终，以所有车辆个体成本之和构造系统社会总成本，并以社会总成本的倒数作为算法适应度函数，从而形成面向新能源汽车共享换电资源分配与多智能体博弈优化算法测试的数据集。
", "ds_quality": "满足相应论文中博弈模型的约束条件。
", "ds_acq_start_time": null, "ds_acq_end_time": null, "ds_acq_place": "", "ds_acq_lon_east": null, "ds_acq_lat_south": null, "ds_acq_lon_west": null, "ds_acq_lat_north": null, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "open-access", "ds_total_size": 13154885, "ds_files_count": 0, "ds_format": "JSON", "ds_space_res": "", "ds_time_res": "", "ds_coordinate": "无", "ds_projection": "", "ds_thumbnail": "de4f5432-c12b-4715-86b8-a180e2ec7cbc.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "结合论文中博弈模型，作为多个模型实例的初始化参数使用。", "ds_from_station": "", "organization_id": "9ecaaa78-39e9-411e-9f24-274e12aa643f", "ds_serv_man": "陈春华", "ds_serv_phone": "15013153445", "ds_serv_mail": "chunhuahcen@scut.edu.cn", "doi_value": "", "subject_codes": [ "410" ], "quality_level": 0, "publish_time": "2026-07-09 10:58:58", "last_updated": "2026-07-09 10:58:58", "protected": false, "protected_to": "2028-06-30 00:00:00", "lang": "zh", "cstr": "", "i18n": { "en": { "title": "Game data set of new energy vehicle power sharing scenarios based on multi-objective complex multi-agent game model based on swarm intelligence evolutionary learning", "ds_format": "JSON", "ds_source": "This dataset consists entirely of simulated generated data and does not originate from any specific literature, actual measurements, or third-party downloads.", "ds_quality": "Satisfy the constraintSatisfy the constraints of the game model in the corresponding paper.", "ds_ref_way": "", "ds_abstract": "The content of the game data set of the new energy vehicle sharing power exchange scenario based on the multi-objective complex multi-agent game model based on swarm intelligence evolution learning is three types of static complete information synchronous game scenarios: small-scale (30 vehicles, 5 power exchange stations), medium-scale (150 vehicles, 25 power exchange stations) and large-scale (600 vehicles, 100 power exchange stations).
\r\nEach class contains 20 independent random instances, for a total of 60 JSON format instance files. Each example records the two-dimensional coordinates of the vehicle and the exchange station (area side length is 20), and the number of batteries at each station (5~10 small/medium scale, 10~15 large scale), initial SoC per battery (0.5~1.0 uniform sampling), initial SoC of vehicle (0.2~0.4 uniform sampling), minimum safe arrival SoC of vehicle (1/3~2/3 of the initial SoC), vehicle power replacement demand threshold (0.5~0.7 uniform sampling) and other parameters, and define individual costs and total social costs based on the dynamic electricity price model.
", "ds_time_res": "", "ds_acq_place": "", "ds_space_res": "", "ds_projection": "", "ds_process_way": "This dataset is constructed through simulation-based generation, with the generation process relying on Python simulation programs and NumPy's random sampling mechanisms. Specifically, the process proceeds as follows. First, according to the experimental requirements of the new energy vehicle (NEV) battery-swapping sharing scenario, different scales of numbers of vehicles awaiting swapping and numbers of swapping stations are configured, yielding three categories of game scenarios: small-scale, medium-scale, and large-scale. The small-scale scenario includes 30 EVs and 5 BSSs, the medium-scale scenario includes 150 EVs and 25 BSSs, and the large-scale scenario includes 600 EVs and 100 BSSs. For each category, 20 independent random instances are generated, resulting in a total of 60 data instances. Subsequently, spatial coordinates of EVs and BSSs are generated within a 20×20 two-dimensional area using random sampling methods. Based on the Euclidean distances between vehicles and swapping stations, the travel distances from vehicles to different swapping stations are computed, providing fundamental data for subsequent calculations of energy consumption, arrival safety constraints, and travel costs. Meanwhile, for each scenario scale, upper and lower bounds on the number of batteries available at a single swapping station are set, and the number of available batteries at each station is sampled via integer uniform sampling. If the total number of generated batteries is less than the number of vehicles, resampling is performed to ensure that the overall resource scale meets the experimental requirements for multi-vehicle swapping decision-making. Furthermore, by incorporating unit-mileage energy consumption coefficients, rated battery capacities, dynamic electricity price parameters, swapping payment cost weights, and travel cost weights, the individual cost for each vehicle under different swapping choices is calculated. For choices that violate local feasibility constraints, a penalty term is added to the cost function. Finally, the system-wide social total cost is constructed as the sum of individual costs over all vehicles, and the reciprocal of the social total cost is used as the algorithm fitness function, thereby forming a dataset designed for testing optimization algorithms for NEV sharing-based battery-swapping resource allocation and multi-agent game problems.", "ds_ref_instruction": "Against the game model in the paper, it is used as initialization parameters for multiple model instances." } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "belong_to_nieer": false, "ds_topic_tags": [ "新能源汽车换电", "博弈模型" ], "ds_subject_tags": [ "工程与技术科学基础学科" ], "ds_class_tags": [], "ds_locus_tags": [], "ds_time_tags": [], "ds_contributors": [ { "true_name": "陈春华", "email": "chunhuahcen@scut.edu.cn", "work_for": "华南理工大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "陈春华", "email": "chunhuahcen@scut.edu.cn", "work_for": "华南理工大学", "country": "中国" } ], "ds_managers": [ { "true_name": "陈春华", "email": "chunhuahcen@scut.edu.cn", "work_for": "华南理工大学", "country": "中国" } ], "category": "其他" }