65results about How to "Solve the problem of cold start" patented technology

The invention relates to a personalized information recommendation method by information fusion, belonging to the technical field of Internet information recommendation. The method includes first pre-processing a data set to extract item shape information: determining a relationship type between items, constructing a relationship network between items, determining item text information, and determining item image information; using a network representation learning method to extract network relationship features, extracting text features by using a text representation learning method, and extracting image features by using an image feature extraction method; then calculating a user's preference feature value for each item in each dimension; finally inputting preference features into a sorting model, and recommending items with scores TOP-N in an alternative set to the user. Compared with the prior art, the method improves the accuracy of recommendation results by excavating and utilizing attribute information of the items to supplement sparse user active interaction data; and the integration of the item attribute information can make the recommendation not only rely on rating data,but also help solve the problem of cold boot of new items.

The invention discloses a method and a device for training a recommendation model, and a recommendation system. The device for training the recommendation model comprises a log acquisition unit, a first characteristic extraction unit, a second characteristic extraction unit, a sample generation unit, a sample aggregation unit and a model training unit, wherein the log acquisition unit is suitable for acquiring a plurality of access logs; the first characteristic extraction unit is suitable for extracting a first characteristic used for identifying recommendation information which presents on a corresponding page of each access log from each access log, and extracting a second characteristic used for identifying the context information of the page; and the second characteristic extraction unit is suitable for acquiring the interest degree sorting value of the user associated with each access log on various pieces of recommendation information, inquiring the interest degree sorting value corresponding to the recommendation information which presents on the corresponding page of the access log, and taking the acquired interest degree sorting value as a third characteristic.

The invention discloses an engine coolant waste heat utilization engine warming system adopting a solid-liquid phase change heat reservoir and a method of the engine coolant waste heat utilization engine warming system. The system comprises a temperature sensor, an electronic thermostat, an electronic fan, a radiator, the solid-liquid phase change heat reservoir, an electronic three-way valve and an MCU. Coolant flows through the phase change heat reservoir to heat phase change materials. The phase change happens to the phase change materials, the phase change materials absorb a large amount of latent heat, and residual heat of an engine is transferred into the phase change heat reservoir to be stored and used for starting engine warming. The electronic thermostat and the electronic three-way valve are controlled by the MCU and kept running synchronously all the time. The premise that normal work of an engine cooling system is not influenced is ensured. After the engine stops, large circulation is closed through the electronic thermostat and the electronic three-way valve. The phase change heat reservoir is sealed and enters the energy preservation state. Thus, under the condition that the existing engine power consumption is hardly increased, waste heat recycle is achieved. The engine warming starting process is achieved under the condition that the engine power consumption is not increased.

The invention relates to a shift register circuit. The shift register circuit comprises a plurality of shift registers for outputting a plurality of drive pulse signals in turn, wherein in M shift registers for outputting M drive pulse signals in turn, the last shift register for outputting the drive pulse signals receives the enabling of (M-1) start pulse signals output by (M-1) shift registers for first outputting the drive pulse signals in turn to generate drive pulse signals, and M is a positive integer and more than 2. In addition, the invention also provides the shift register.

The invention is applicable to the technical field of computers, and provides an article recommendation method and a device. The method comprises the following steps of according to the similarity between a scored article and a to-be-scored article in historical scoring data, calculating the influence factor of the historical scoring data on the predicted score of the to-be-scored article; inputting the calculated influence factor into a pre-established restricted boltzmann machine model; calculating the predicted score of a user on each to-be-recommended article through the restricted boltzmann machine model; generating a recommendation list; and outputting recommended articles to the user according to the generated recommendation list. In this way, the cold start problem during the new article recommending process can be solved. The recommendation accuracy is improved.

The invention relates to a video recommendation method, a video recommendation device and a computer readable storage medium. The video recommendation method comprises the steps of obtaining an equipment ID of an unregistered user; Retrieving friend relationship data according to the device ID to obtain a first user ID corresponding to the device ID; Sending the video watched by the first user IDas a recommended video to a user of the device ID, the step of obtaining friend relationship data comprising: when an unregistered user watches the video based on a shared video link, obtaining the first user ID from the video link; And establishing an association relationship between the device ID of the unregistered user and the first user ID. According to the video recommendation method, the unregistered user watches the video shared by the registered user, so that the association relationship between the equipment ID of the unregistered user and the user ID of the registered user is established, the video is recommended to the unregistered user who watches less historical data accordingly, and the problem of cold start of the user is solved.

The invention discloses an online commodity recommendation method based on user novelty. The method comprises the following steps: S1, forming a selection sequence chi according to commodity browsinginformation of a user u; s2, determining the window time Wt of the user according to the click frequency of the user u; s3, obtaining the user novelty through calculation, wherein the probability thatthe user u selects a new commodity at the moment t is predicted according to the behavior historical record of the user u when the user selects to start to cut off the moment t; and S4, according tothe obtained user novelty, recommending commodities or commodity combinations meeting the user novelty at the moment t to the user. According to the online commodity recommendation method, the user novelty is calculated by introducing psychological indicator factors of the user for the openness degree of new and old commodities, and commodity / commodity combination recommendation is performed on the user according to the user novelty, so that the online commodity recommendation method can improve the accuracy of personalized recommendation, and more economic benefits are created for an e-commerce platform.

The invention is suitable for the technical field of information and provides a method and a device for recommending items. The method comprises the following steps: acquiring item information within a preset distance scope on a present position of a to-be-recommended user; acquiring a candidate user group according to the item information; calculating an interactive behavior similarity and an item scoring similarity between the to-be-recommended user and each candidate user in the candidate user group; calculating a compound similarity between the to-be-recommended user and each candidate user according to the interactive behavior similarity and the item scoring similarity; selecting N candidate users with the maximal compound similarity and forming a neighbor set of the to-be-recommended user; acquiring the item concerned by each candidate user in the neighbor set; and forecasting a score of the item scored by the to-be-recommended user and recommending M items in the highest scores to the to-be-recommended user. According to the invention, the problems of sparse data and cold starting of the present individual recommendation technique can be solved and the item recommendation accuracy is promoted.

The invention provides a resource recommendation method and device. The method comprises the following steps of: respectively inputting a marketing resource training sample and a customer resource training sample into a resource classification model to carry out training so as to obtain a marketing resource classification result and a customer resource classification result; and inputting the marketing resource classification result and the customer resource classification result into a similarity model to carry out training so as to output a plurality of matching results as recommendations. According to the resource recommendation method and device, on one hand, modeling training is carried out through machine learning to intelligently classify marketing resources and customer resources,and the classification results are intelligent matched, so that the problem of cold starting is solved, and more accurate recommendation relative to existing scheme is realized; and on the other hand,a large amount of data information related to customers or potential customers is mined through intelligently classifying customer resources, so that data assets are created while CRM accurate recommendation is completed.

The invention discloses a museum cultural relic guide and knowledge recommendation method based on a knowledge graph. The method comprises the steps that firstly, a word segmentation tool and a regular expression conversion rule are used for converting natural language query input by a user into SPARQL query statements, cultural relic knowledge query and reasoning are carried out in a SPARQL language mode, and the queried cultural relics are visually presented to the user in a knowledge graph mode; meanwhile, through recommendation algorithms such as cultural relic attribute association and collaborative filtering, cultural relics or cultural relic background knowledge related to currently inquired cultural relics are recommended to the user for the interested user to select and browse ina knowledge graph mode. The method can interact with the user through the natural language based on the constructed cultural relic knowledge map, recommend other cultural relics related to the currentcultural relic or possibly interested by the user to the user, guide the user to better browse cultural relic information and understand related cultural relic knowledge.

The invention relates to a data anomaly detection method and device, a readable storage medium and an electronic device. The method comprises steps as follows: acquiring current index data; acquiring a time sequence set corresponding to the current index data; as for each time sequence in the time sequence set, whether a memory neuron for recording the time sequence exists is judged; when the memory neuron is judged to exist, the memory neuron is activated; when the memory neuron is judged not to exist, one memory neuron is allocated for the time sequence to record the time sequence, and the allocated memory neuron is activated; whether the current index data are abnormal is determined at least according to the activated memory neuron. Accordingly, the anomaly monitoring system has higher adaptability to different types of index data and complex change index data, the larger the scale of the acquired index data is, the higher the study capability and judgement capability is, and the higher the accuracy of anomaly detection is.

The invention relates to methanol gasoline for vehicles and a preparation method thereof, belonging to the technical field of the petrochemical industry. The methanol gasoline for vehicles is prepared from the following raw materials in percentage by weight: 70-85 percent of gasoline, 10-25 percent of methanol, 0.5-8 percent of petroleum ether, 0.1-2.5 percent of ferrocene, 2,2,3',3'-tetramethylbutane, 0.1-2 percent of di-tert-butyl, 0.5-6 percent of methyl tert-butyl ether, and the like. The preparation method comprises the following steps of: preparing an additive; preparing denatured methanol; and mixing with gasoline and stirring to obtain methanol gasoline for vehicles. By adding some common reagents, the prepared methanol gasoline for vehicles is suitable for various vehicles, has favorable cold start without depositing carbon and clean and environment-friendly exhaust emission without changing the vehicle structure, can be used singly or used with various current brands of gasoline and has favorable economy and market prospects.

The invention provides a game character creating method, a game character creating device, electronic equipment and a computer readable storage medium, and belongs to the technical field of program development. The method includes the steps: acquiring an initial attribute of a character to be built; determining a similar character of the character to be built in existing game character according to the initial attribute, and extracting configuration data of the similar character; applying the configuration data of the similar character to the character to be built, and performing simulation game to obtain simulation game data of the character to be built; adjusting the initial attribute of the character to be built according to the simulation game data to obtain a final attribute of the character to be built. The configuration data of the similar character at least comprises equipment data of the similar character. The method can solve the problem of cold starting of the configurationdata in creating of the game character, creating processes of the game character are simplified, and efficiency is improved.

The invention provides a function recommendation method applied to electronic equipment. The method comprises the steps: obtaining first user information of a first user, wherein the first user information comprises first user feature information; generating a first feature vector according to the first user information and the knowledge graph; processing the first feature vector by adopting a neural network model, and determining scores of the plurality of functions for the first user; and determining a function to be recommended to the first user in the plurality of functions according to the scores of the plurality of functions for the first user, wherein the knowledge graph comprises nodes for the first user characteristic information and nodes for multiple functions, the nodes for thefirst user characteristic information point to the nodes for the first user characteristic information. The invention further provides a function recommendation device applied to the electronic equipment, the electronic equipment and a computer readable storage medium.

The invention discloses a sequence recommendation method and system based on federal learning, and belongs to the field of personalized recommendation and the field of user privacy protection. According to the method, personal information and real-time and historical data of a user are stored locally, each client can perform state representation according to own historical data, an interaction relationship between the user and an article is captured, and a recommendation process is regarded as a serialized decision process based on deep reinforcement learning; according to the method, an attention mechanism is introduced into an aggregation algorithm of federated learning, correlation of feature information extracted by each client and data difference brought by individuation of each client are considered, a weight coefficient is formulated for each client, which is a fine-grained re-weighting means, the individuation degree of data is increased, and the robustness of the data is improved. According to the method, the recommendation accuracy is improved, the method is more suitable for the field of sequence recommendation, and personalized recommendation for the user is realized while personal data privacy is protected.

The invention discloses a big data based topic selection planning system. The big data based topic selection planning system includes a data acquisition unit for acquiring associated data of book topic selection planning; a data processing and storage unit for dividing the acquired data into a real-time service data source and a basic data source according to the data timeliness demand, storing the data in a hadoop based publishing house distributed database, and uploading the date into a data mart; an analysis prediction unit including a statistic analysis platform, wherein the statistic analysis platform can perform overall analysis and prediction on the data provided by the data mart, and can provide an important reference and basis for making and modifying of topic selection planning of a publishing house. The big data based topic selection planning system uses a big data analysis technique to help a book publishing mechanism to master book market purchase trend and market hot spots, can timely update the most suitable topic ranking, can accurately master the actual demands of a reader group, can position a specific consumer group, and can provide a scientific basis for improvement of topic selection planning ability of the book publishing mechanism.

The invention discloses methanol petrol for vehicles, and belongs to the field of fuel. The methanol petrol is prepared from the following raw materials in parts by weight: 65-85 parts of methanol, 5-12 parts of petrol, 0.8-1.5 parts of paraffin oil, 2-6 parts of glycerine, 1-5 parts of acetone, 1-5 parts of acidamide, 0.8-3 parts of tertiary butyl acetate, 1-4 parts of dimethyl phthalate dibutyl, 0.5-1.5 parts of glyceryl monostearate, 1-5 parts of ethybenzene, 1-4 parts of tert-butyl alcohol, and 0.2-0.5 part of iron chloride. The oxygen content of the methanol in the methanol petrol for the vehicles is high, the methanol petrol is sufficiently combusted, black smoke is not generated, carbon deposits are not generated, the bottom of a pan is not sooted, raffinate and residues are not generated, the exhaust emission amount after the combustion of the methanol petrol is low, the methanol petrol is clean and environment-friendly, the product performance is stable, the calorific value is high, the cost is low, the methanol petrol is safe and reliable, an engine does not need to be modified, the methanol petrol can directly replace petrol, and the methanol petrol has a broad application scope.

The invention discloses an API recommendation method based on a knowledge graph and collaborative filtering. The method comprises the following steps: constructing a service knowledge graph; embedding API entities in the knowledge graph into a low-dimensional space, and calculating the similarity Sim1 between APIs; obtaining APIs used by the target Mashup, and enabling similar APIs of the APIs to form a recommendation set RS1 based on Sim1; extracting functions between the Mashups according to the Mashup description, and calculating the similarity Sim2 between the target Mashup and other Mashups through the extracted functions; obtaining a similar Mashup of the target Mashup based on the Sim2, and combining APIs used by the similar Mashup into a recommendation set RS2; constructing a usage matrix of the Mashup and a usage matrix of the API based on a historical usage relationship between the Mashup and the API, and calculating the similarity Sim3 between the Mashup matrixes and the similarity Sim4 between the API matrixes; obtaining a similar Mashup of the target Mashup based on the Sim3, and enabling APIs used by the similar Mashup to form a recommendation list RS3; obtaining APIs used by the target Mashup, and enabling similar APIs of the APIs to form a recommendation list RS4 based on Sim4; and finally, obtaining a final API recommendation result according to the RS1, the RS2, the RS3 and the RS4.

The embodiments of the present invention disclose a user similarity-based cross-domain scoring method and a scoring device. The invention aims to solve the problem of cold start of potential users ina target domain. The method provided by the embodiments of the present invention includes the following steps that: a user-item scoring matrix in a target domain and a user-item scoring matrix in an auxiliary domain are obtained, users in the target domain are included in users in the auxiliary domain; the values of similarity between the users in the auxiliary domain are obtained according to theuser-item scoring matrix of the auxiliary domain; and the predicted scores of the users in the auxiliary domain for corresponding target items in the target domain are calculated according to the values of similarity between the users in the auxiliary domain and the user-item scoring matrix in the target domain, wherein the target items are items which are not scored in the target domain by the users in the auxiliary domain. The embodiments of the present invention further provide a scoring device for solving the problem of the cold start of potential users in the target domain.

The invention relates to a methanol gasoline additive for vehicles and a preparation method thereof, belonging to the technical field of petrochemical industry and aiming to solve the problems of mutual dissolution, corrosion, air resistance, swelling and the like in the preparation process of methanol gasoline. In a technical scheme, the methanol gasoline additive for vehicles is prepared by mixing the following raw materials in parts by weight: 15 to 30 parts of ethanol, 10 to 20 parts of isopropyl ether, 1 to 9 parts of dimethyl carbonate, 1 to 15 parts of acetone, 0.1 to 6 parts of dimethyl ketone, 0.5 to 8 parts of isopropanol, 1.5 to 10 parts of tertiary butanol and 1 to 8 parts of iron naphthenate. The preparation method comprises the steps of weighing raw material components according to the weight parts, and mixing and standing under normal temperature and pressure. The methanol gasoline additive for vehicles has the advantages of good safety and environment protection, and well solves the problems of mutual dissolution, corrosion, air resistance, swelling and the like in the preparation process of the methanol gasoline, and the preparation method has the advantages of preparation under normal temperature and pressure, simple technology, and simple and convenient operation. The methanol gasoline blended with the additive is suitable for various vehicles and has excellent economy and market application prospects.

The invention discloses alcohol-based fuel without eye fumigation in use. The alcohol-based fuel is characterized by being prepared from the following ingredients in parts by weight: 85.0 to 95.0 parts of methyl alcohol, 2.0 to 8.0 parts of ethyl alcohol, 2.0 to 5.0 parts of tertiary butanol, 1.0 to 2.0 parts of paroline, 0.2 to 0.5 part of ferric chloride, 0.2 to 0.5 part of methyl tertiary butylether, 0.05 to 0.1 part of ethyl acetate and 1.5 to 2.5 parts of additive. The additive is prepared from the following preparation method which comprises the steps: (1) forcing treatment: putting water-extracted folium eriobotryae residues into a high-pressure tank, forcing to 3kg / cm<2> to enable pressure to be even, maintaining and then releasing the pressure instantly; (2) distillation treatment: putting the materials obtained in the step (1) into a distiller, heating, burning or introducing steam at the bottom of the distiller, guiding the steam into a condenser through a condensing pipe when the distiller is full of the hot steam and then collecting an oil phase through an oil-water separator; (3) high-shear homogenizer treatment: processing the oil phase obtained in the step (2) by ahigh-shear homogenizer to be prepared into nanoemulsion to obtain the additive.

The invention discloses a commodity matching recommendation method and a commodity matching recommendation device based on attributes and titles. The commodity matching recommendation method comprisesthe following steps of: acquiring matching scores among commodities by acquiring user order information within a period of time, acquiring a candidate commodity set with a relatively high matching score according to the condition of the matching score, and introducing heterogeneous co-occurrence frequency and heterogeneous co-occurrence inverse frequency to reduce the influence of hot commoditieson commodity matching, so as to obtain the relatively accurate candidate commodity set; meanwhile, acquiring vectors of attributes such as commodity styles and vectors of commodity titles and mappingthe two types of vectors into a vector space, learning the matching degree between the two commodities, establishing a commodity matching model, and then acquiring a matching rate between the commodities more accurately based on commodity attributes and titles; and selecting a candidate commodity which can be matched with the highest matching degree from the candidate commodity set of the commodities according to the commodity information purchased or inquired by the user. The commodity matching recommendation device is based on the commodity matching recommendation method, and aims to precisely recommend the matched commodities.

The invention discloses an electric automobile battery box based on a liquid-cooled loop and active heating. An upper battery pack, an upper mounting plate, a lower battery pack and a lower mounting plate are sequentially arranged in a battery box shell from top to bottom, a cooler is mounted on the outer bottom surface of the battery box shell, heat conducting fillers are mounted between the upper battery pack and the upper mounting plate, between the lower battery pack and the lower mounting plate, between the lower mounting plate and the inner bottom surface of the battery box shell and between the cooler and the outer bottom surface of the battery box shell, a liquid-cooled pump and a liquid storage device are communicated through a liquid pipeline and mounted in the battery box shell, the liquid pipeline is sequentially connected with the liquid-cooled pump, penetrates the upper mounting plate and the lower mounting plate, penetrates out of the battery box shell, penetrates the top mounting surface of the cooler and detours, penetrates into the battery box shell and is connected with the liquid-cooled pump to form a loop. According to the electric automobile battery box, heat of electric automobile batteries is discharged by the aid of a circular liquid-cooled loop, and heating and temperature control of the battery packs are performed by the aid of a thin-film electric heater or an electric heating wire when the temperature is low, so that low-temperature start is achieved.

The invention relates to the technical field of biomass fuel, and particularly relates to biomass synthetic fuel. The biomass synthetic fuel comprises the following components in parts by weight: 80-95 parts of an alcohol, 30-50 parts of an alkane, 10-20 parts of diethyl ether, 4-8 parts of a combustion-supporting heat-increasing agent and 0.5-4 parts of a safety stabilizer. The beneficial effectsof the biomass synthetic fuel provided by the invention are that the alkane is added on the basis of an alcohol group, so that the overall combustibility is improved; and the combustion-supporting heat-increasing agent and the safety stabilizer are added, so that the safety of the fuel is improved, the heat value of combustion is improved, and the flame temperature is increased.

The invention relates to the field of fuel preparation, particularly to a novel liquid fuel and a preparation method thereof. The novel liquid fuel provided by the invention is prepared from the compounding of a main material and three kinds of auxiliary solvents, the auxiliary solvents are successively mixed, the obtained mixed auxiliary solvent has good intersolubility, and no other impurities are generated; the mixed and dissolved auxiliary solvent is added in the main material so as to mutually dissolve with the ingredients in the main material, as a result, the ingredients can sufficiently react, the ingredients cooperate with one another, the prepared novel liquid fuel effectively solves the problem that the octane values of most standard fuels are substandard, and the anti-detonating quality enhancing effect is excellent; the problem of cold ignition boot is solved; the novel liquid fuel has the characteristics of energy conservation, emission reduction, high effects, environmental protection and easiness in use, the novel liquid fuel can be singly used in a large amount for a long time, and also can be mixed with conventional commercial gasoline, such as number 97, 93, in any proportion, bad influences cannot be generated, engines and the like do not need to be washed, and the novel liquid fuel has relieving and supplementing effects for the current domestic situation of resource shortage to some extent.

The invention belongs to the field of liquid fuel cells. The invention relates to a method for low-temperature fast start-up of a liquid fuel cell system. The method is characterized in that organic small molecular substances such as methanol and ethanol are used as fuel and catalytically combusted in a catalytic combustor, and the heat generated by combustion is directly or indirectly used to heat an electric pile through an electric pile heating chamber and / or to heat a water and heat management component in the system, so as to quickly start the liquid fuel cell system at low temperature. Compared with the prior art, the method helps to start the liquid fuel cell system at low temperature, improve the low-temperature environment adaptability of liquid fuel cells and expand the application scope of liquid fuel cells.

The invention relates to a micro-service management method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a running version and an effective version currently corresponding to an interface of the micro-service; the effective version is a to-be-operated version determined based on an operation version updating request sent by a terminal; when the running version is inconsistent with the effective version, obtaining interface code data corresponding to the effective version; and under the condition of keeping the online state of the micro service, running the interface code data to update the interface. By adopting the method, the hot update of the interface can be realized under the condition that the service residence is ensured, so that the cold start problem of the micro-service is solved, and the processing efficiency of the micro-service can be improved.

The invention discloses a personalized video recommendation method and system, and the method comprises the steps: A, carrying out the preprocessing of all current videos, and extracting the codes ofthe videos; b, for a new user, randomly recommending a batch of videos; c, calculating the fitness of each recommended video to the user according to the operation feedback of the user on the batch ofvideos, including like clicking, commenting, forwarding, watching length, watching times and other factors; d, preferentially selecting the previously recommended video through the fitness; e, randomly enabling the previously selected video codes to generate some mutations, performing some mutations, and giving some new attempts to the user; according to the method adopted by the invention, shortvideo recommendation can be carried out on the user to a certain extent, a certain learning ability and interpretability of a result are achieved, and the problem of cold start can be solved to a certain extent.

The invention belongs to the technical field of graph neural network recommendation, and discloses a graph neural network recommendation method and system fusing a social relation and a semantic relation, which utilize a simple graph neural network recommendation model to recommend a graph neural network based on user social information, score interaction information of users and items and label information of the items. Constructing the information of the non-graph structure into graph information; clustering the labels of the items by using the label information of the items; and calculating the relevancy between the items and the label clusters, carrying out cosine similarity calculation by utilizing an item-label cluster relevancy matrix, and when the similarity is greater than a certain threshold value, adding a trust relationship between the two items to form a semantic relationship graph structure of the items. According to the method, the user-project score bipartite graph is used as a basis, other isomorphic graphs are used as supplements, so that the purpose of using more neighbor nodes for aggregation is achieved, the problem of data sparsity is solved, and tests prove that the proposed model has higher recommendation accuracy than GrpahRec.

The invention discloses an engine coolant waste heat utilization engine warming system adopting a solid-liquid phase change heat reservoir and a method of the engine coolant waste heat utilization engine warming system. The system comprises a temperature sensor, an electronic thermostat, an electronic fan, a radiator, the solid-liquid phase change heat reservoir, an electronic three-way valve and an MCU. Coolant flows through the phase change heat reservoir to heat phase change materials. The phase change happens to the phase change materials, the phase change materials absorb a large amount of latent heat, and residual heat of an engine is transferred into the phase change heat reservoir to be stored and used for starting engine warming. The electronic thermostat and the electronic three-way valve are controlled by the MCU and kept running synchronously all the time. The premise that normal work of an engine cooling system is not influenced is ensured. After the engine stops, large circulation is closed through the electronic thermostat and the electronic three-way valve. The phase change heat reservoir is sealed and enters the energy preservation state. Thus, under the condition that the existing engine power consumption is hardly increased, waste heat recycle is achieved. The engine warming starting process is achieved under the condition that the engine power consumption is not increased.