Keywords: intelligent medical terminal; data utilization; TOPSIS evaluation model; Dijkstra's optimal path algorithm
I. Introduction
The complexity of road conditions in large cities, relatively congested roads, the relative concentration of comprehensive treatment hospitals and specialized hospitals, and the large number of foreign patients coming to hospitals make the hospitals overcrowded and have an impact on the traffic around the hospitals. For patients, time is of the essence. For patients, time is life, in the case of traffic accidents or sudden accidents at home, timely access to medical care is particularly important.
Path selection is an important part of this system, and its study has been well established in the literature both at home and abroad. Japan launched the CACS program in 1973, which developed a dynamic path guidance system for vehicles based on RF radio frequency communication, which can reduce travel time by 13%. The U.S. introduced two typical autonomous path guidance systems, TravTek and Advance, for dynamic path selection for vehicles. Germany developed the LISB system and AutoGuide system based on infrared beacon communication in the 1980s, while later on the world's first commercial on-board path guidance system, Tramc Master, was introduced in the UK [1]. Meanwhile, there is also Eck et al. conducted a study on the problem of pharmacy location selection based on accessibility using GIS technology as an analytical tool [2].
Wu et al. analyzed the spatial accessibility of rural medical facilities with Lankao County, Henan Province as an example [3]. Xiong Juan et al. analyzed the equalization of medical services in Songzi City, Hubei Province based on accessibility [4]. Zhang Li et al. developed a path selection information system based on the shortest time to evaluate the accessibility of hospitals in Yizheng City, Jiangsu Province [5].
The above studies only considered road accessibility from the perspective of transportation, while the smart medical terminal application model based on Internet+ proposed in this paper, on the basis of road accessibility, combined with the Internet open data, combined the use of transportation data and medical data, introduced the internal factors of hospitals, such as hospital traffic flow and other internal factors and factors such as the needs of patients' conditions, and constructed a multi-indicator evaluation based on the selection of goals and the corresponding model of path planning. With the support of the Internet, the system can provide people with faster and more comfortable medical assistance services, and to a certain extent, guide the rational allocation of medical resources, a large amount of data at the same time also provides guiding suggestions for efficient decision-making by government management [6].
Second, the establishment of intelligent medical terminal application model
(a) the overall design of the model
In the era of the Internet + era, access to information on the road conditions of the city, the future can be accessed in the city hospitals (medical institutions) real-time outpatient in the number of people in the hospital, real-time number of people waiting in each department of the hospitals, the number of hospitals weekly appointments for specialists, and other medical information. An application model is constructed based on the above data. This paper develops a simulation system based on this scheme to realize the informatization, rapidity and comfort of patient consultation [7]. The block diagram of the overall program is as follows:
(II) Road network structure and hospital setup
1. Taking part of the road sections in Binhu District, Beitang District, Chong'an District, and Nanchang District of Wuxi City as a prototype, the urban road network is constructed. The road network is configured with different levels of roads, which are urban expressway and ordinary urban roads. According to the road traffic safety regulations, the model is set in the car in the city expressway driving average speed of 40km / h, in the average speed of ordinary urban roads for 28km / h.
2. Roads in the city may be congested, and the probability of congestion in the city center as the origin of the radiation to the surroundings to reduce the model algorithms to automatically choose to bypass the section of the road once the congestion occurs.
3. Selected four hospitals as this system set the hospital point, as shown in the box below, respectively, given the virtual grade Wuxi People's Hospital (three-star), Wuxi 101 Hospital (two-star), Taihu Street Health Service Center (one-star), Binhu Xuelang Community Health Service Center (one-star).
(C) Establishment of optimal path evaluation model
For the given four hospitals, the final score is weighted by three parts. They are journey time score, hospital real-time congestion score, and hospital grade score. Higher individual scores indicate that patients are more inclined to choose that hospital. The model mainly contains the shortest path problem and the evaluation problem. Among them, the travel time score, the hospital real-time congestion two scores using TOPSIS method for analysis and calculation.
The TOPSIS method is a decision-making method commonly used for multi-objective decision analysis of limited programs in systems engineering. It identifies the optimal and worst solutions in the finite program from the normalized raw data, and then finds out the relative proximity between the evaluation object and the optimal and worst solutions through the distance between the evaluation object and the optimal and worst solutions, which serves as the basis for comprehensive evaluation. The method is characterized by simple calculation, reasonable results and flexible application [8].
1. Journey Time Score
Since the passage is time-consuming, the amount of journey can be transformed into the amount of time, i.e. the shortest journey corresponds to the shortest time. In this model, the distance of all road sections is expressed as the adjacency matrix A, A(i,j) denotes the length of the road section ij, which is set to be infinite if there is no road section connectivity. The optimal path is calculated using the Dijkstra single-source shortest path algorithm, i.e., using the adjacency matrix. The Dijkstra algorithm for solving the shortest path between two specified vertices u0 and v0 is shown below as a flowchart of the algorithm. at the end of the Dijkstar algorithm, the shortest distance from u0 to v0 is given by L(v0) and where d(u0,v0) denotes the distance between two points [9].
The different road classes correspond to different average speeds, and the change in speed is converted into a change in distance in order to modify the distance matrix (introduced in the design of the simulation system), and the time consumed to pass on an ordinary urban road is twice as long as that of an urban expressway, so the distance matrix is updated accordingly, and the updated matrix is used in the actual computation of the travel time scores while the original distance matrix is still used in calculating travel time values. Neighborhood matrix.
A new adjacency input matrix is formed by adding congestion factors, road class and other factors to urban roads. The shortest distance to each of the four hospitals can be found by the shortest path algorithm. Because in the scoring, the need to ensure the consistency of the scoring scale, that is, the first to solve the shortest distance to all hospitals, and then to the four shortest distance as a comparative object, the "best of the best". The four shortest distances are
The distance needs to be TOPSIS normalized, with the following formula:
The value is located between [0,1], because of the agreement that the higher the score, the more optimal the path, the
And at that time, the score is 0. The score will be corrected to
The formula calculates the score, there will be no score of 0, and also have a good distinction between the four hospitals. differentiation.
2. Hospital Congestion Score
Real-time outpatient in-patient number = the number of hospitals have been registered on the day - the number of outpatient visits completed.
The greater the number of outpatients in the hospital, often means that the longer the waiting time for registration, examination, treatment, this paper uses the number of real-time outpatient hospitals in the approximation of the number of people who need to wait in line for medical treatment as a measure of the length of time. And the ratio of the number of outpatients in the hospital to the number of people in the hospital at capacity indicates the real-time degree of congestion in the hospital.
The formula for crowding in each hospital is expressed as follows:
Rate value is a number between 0 and 1, and the formula for the crowding score after TOPSIS normalization is:
3. Scale", set three-star, two-star, one-star hospital scores as shown in Table 2:
4. Calculation of route score formula
Taking into account the different needs of each patient, the system in the path planning set the system recommended the best, the shortest passage time, the best hospital grade, the hospital queuing the shortest priority of the four path planning preferences. Preferences. The length of the journey corresponds to the travel time, the saturation of the flow of people corresponds to the degree of congestion, and the different star ratings of the hospitals correspond to the hospital grade scores. Finally, the final score is weighted according to a certain weight to obtain a weighted average, inverse its optimal path program, to solve the optimal program driving distance and time consumed.
5. Real-time hospital statistics
The model assumes that the real-time number of people in the hospital outpatient clinic can be obtained (the number of people in the hospital = the number of people who have registered today - the number of people who have completed the consultation), the number of people waiting in real time for each department, the number of appointments last week Top5, and weekly traffic flow of the hospitals. Each of the above data can be visualized. The weekly appointment volume of the specialist outpatient clinic may be showing cyclical changes, so the data is fitted with a cubic polynomial to present a trend chart of the flow of people, which can be used to predict the future flow of people.
The fitting curve uses the cubic curve least squares method, because the cubic curve has at most two extreme points, and its trend can satisfy the description of the trend of change in the seven days of the week. It is calculated by assuming that the cubic curve is
The least squares algorithm is used to find the vector a that minimizes the sum of the squares of the distances from the points on the curve to the points of the true value.
Third, the system simulation design
1. Hospital and road simulation
(1) Combined with Figure 2, analyze the construction of the road network structure, and mark the length of each section of the road, in order to facilitate the calculation of the shortest distance. Simulation system, the construction of a city with v, b, w, x four hospitals (hereinafter are replaced by the node number), corresponding to Figure 4, four points labeled in blue, the other nodes are crossroads. It is assumed that all nodes except hospitals can be used as user departure points, as shown in Figure 4.
Assume that the area where v-m is located is the city center. Based on the city center and the main roads nearby may be congested, set part of the road with different probability of congestion. Through the random number function to generate a random number between 0 ~ 1, set to x, by judging x belongs to the range of intervals (the length of the interval that is the probability of congestion on the road) to determine for a section of congestion. Such as: x for 0.3 ~ 0.35 when the section kl congestion, and x = 0.31, can be determined for the kl section congestion.
(2) road level setting: the system sets up urban expressway and ordinary city road according to certain road level matching, and sets up hb, li, ic, mv, jd, rs, sx road sections as ordinary city roads (including rural roads and old roads in the main city), and the average traveling speed of the above seven road sections is set up as 28km/h, and the rest of the road sections are set up as 40km/h.
2. Hospital congestion simulation realization
In the system using the random generation of traffic method, first of all need to determine the base of the hospital traffic and saturation traffic, because it is a ratio problem, in the weighting of the score is only taken into account, so the oversaturation situation is ignored here.
According to the different hospital levels of the four hospitals are assigned as shown in Table 2, the number of people base and saturation number:
The system uses simulation of real-time monitoring of the flow of people in MATLAB to set a timer, through the stochastic function to generate a value based on the base change, so as to achieve the effect of simulation. Simulation of the flow of people formula:
The timer is triggered every time, the flow of people is updated once.
3. Route composite score simulation implementation
The system in the path planning set the system recommended the best, the passage of the shortest time-consuming, the hospital level is optimal, the hospital queue waiting for the shortest priority of the four path planning preferences, which need to be considered to generate four groups of different weights to meet the needs of different users.
The weights are assigned as follows:
4. Simulation of hospital statistics
The above congestion data simulation method is still used to set the base value of the flow of different levels of hospitals and the range of change, due to space constraints, only the base value of the weekly flow of different hospitals is listed here.
The simulation effect diagram of hospital traffic statistics is as follows:
Fourth, the model evaluation and simulation test
The smart medical terminal application model for the first time will be obtained from the road data, real-time traffic flow in hospitals, hospitals, hospitals, and other factors into account, combined with the user's preferences and TOPSIS evaluation model, to solve the best hospitals and the best paths to reach the hospitals from the current location. the best path to that hospital. In the system, the effect of different levels of highway on traveling speed is added, and then the effect of speed is converted into the change of distance. When filtering hospitals, the system assigns different weights to different path influences based on patient preferences. At the same time, the system also adds the analysis of the real-time flow of people in the hospital, so as to facilitate the patients to choose the appropriate amount of time to seek medical treatment, in order to reduce the waiting time in the queue. In summary, the system integrates the possible situations on the way to the doctor and after arriving at the hospital, and has good feasibility.
The flowchart of the simulation system operation is as follows:
(1) Record of a single experiment
As shown in the above figure, when taking q as the starting point and selecting the system recommendation, the system gives the best hospital as the People's Hospital of Wuxi City (v) according to the set algorithm, and the best path is q-l-m-v, with the shortest distance of 13.34km. this distance is relative to the remaining three hospitals ( Hospital b: 15.49km, Hospital w: 15.54km, Hospital x: 18.02km) is the smallest, and Hospital v is the only three-star hospital. After determining the hospital and the path the system will automatically mark the path on the map with a green line [10].
(2) Record of multiple experiments
Test the distribution of the total score points: the three starting points of a, p and q are experimented with different path planning preferences and record the score of each hospital and the optimal hospital number each time. The optimal score value recommended by the selection system is recorded as follows:
And for different path planning preferences, the optimal hospital selection results are as follows:
V. CONCLUSION
The intelligent medical terminal application model proposed in this paper, based on the road data, real-time hospital traffic, and basic data of the hospital level, will be the flow of people, the journey time consumed through the TOPSIS evaluation method to be normalization, which makes the scores have scale consistency. Comprehensive time-consuming, hospital congestion, hospital grade three scores, the user path planning preferences into three score weights, calculate the total score, so as to get the best hospital and plan the best path for medical treatment from the current location, the user medical travel more convenient. In the simulation system, the road network structure of Wuxi City is directly used to introduce the impact of different levels of highways on the travel time, different random probabilities are assigned to generate congested road sections and detours according to the real road conditions, and hospital traffic flow is dynamically generated within a certain range according to different levels of hospitals in a random number manner, which is of realistic value in the simulation. At the same time in the model to add the relevant hospital historical data and its function fitting, in order to get the trend of data changes.
The above simulation results show that the single experiment results in a reasonable path planning, in line with the general user needs. Multiple experiments in the record (shown in Table 5), the user is at point a, using the model evaluation program for the four hospitals scoring results were 69.63, 91.75, 65.67, 46.33, different hospitals scored a reasonable distance between the scores are ladder distribution, has a strong degree of differentiation. When the user is at point a,p,q respectively, the recommended hospitals meet the user's needs as shown in the map. The standard deviation of the scores of the four hospitals in the three sets of data from different starting points is 18.63, 13.73, 14.66, and the degree of dispersion is close to the same, i.e., each time the evaluation algorithm is executed, the scores are consistent, and the model is robust. When selecting hospitals with low congestion priority, the system will recommend w hospitals (i.e., community hospitals), which can divert patients with less serious and urgent medical conditions from large hospitals, thus promoting the rational allocation of medical resources and the implementation of hierarchical diagnosis and treatment. Combined with the map and software calculations, the medical coverage of the area (Wuxi urban area) can be determined.
References
[1] Xia Limin. Research on optimal path selection algorithm in transportation system [D]. Capital Normal University, 2007.
[2] Van Eck J R, De Jong T. Accessibility analysis and spatial competition effects in the context of GIS-supported service location planning. Computers, Environment and Urban System, 1999, 23(2): 75-89.
[3] Wu JJ,Kong YF,Li B. GIS-based spatial accessibility analysis of rural medical facilities: a case study of Lankao County, Henan Province. Human Geography, 2008, 23(5): 3742.
[4]Xiong Juan,Luo Jing,Peng Jing,et al. Analysis of county medical service equalization based on accessibility: the case of Songci City in Hubei Province. Human Geography, 2012, 27(5): 2529.
[5]ZHANG Li,LU Yuqi,ZHAO Yuanzheng. Evaluation and Planning of Hospital Accessibility--Taking Yizheng City of Jiangsu Province as an Example[J]. Human Geography, 2008(2):60-66.
[6]XU Shofeng,QIAO Yali,YANG Bin,et al. Discussion on the construction of intelligent medical care in the era of big data[J]. Medical Information, 2015(15):1-1.
[7]Chang Chaodi,Chen Min. Research on healthcare data governance methods in the era of big data[J]. China Digital Medicine, 2016, 11(9):2-5.
[8]Cheng Liling,Cao Jian. Application of weighted TOPSIS method in comprehensive evaluation of hospitals[J]. China Hospital Statistics, 2006, 13(1):17-19.
[9]Bao Peiming. Optimization of Dijkstra's algorithm in distance seeking[J]. Computer Research and Development, 2001, 38(3):307-311.
[10]ZHOU Xinming,LAN Sai,XU Yan. Comparison of several edge detection algorithms in image processing[J]. Modern Power, 2000, 17(3):65-69.