Research on Innovative Strategies of Artificial Intelligence Technology in Smart City Management and Its Practical Effects

In recent years, with the continuous expansion of urban scale and the dramatic increase of urban population, the task of urban management is arduous. The complexity and vulnerability of cities and the uncertainty of emergencies have put forward new requirements for urban management [1-2]. Urban management should be as fine as embroidery, and exploring a new way of social governance that meets the characteristics and development of the city is a big issue for the long-term development of the city [3-4]. Smart city management must be adapted to local urban development. Continuous efforts and innovations should be made to improve the social governance capacity and enhance the vitality of social development. Relevant government documents point out that it is necessary to strengthen the combination of artificial intelligence and social governance, develop artificial intelligence systems suitable for government services and decision-making, and use artificial intelligence to improve the government’s public service capacity and social governance level [5-8]. With the support of government policy, artificial intelligence has become a new engine for urban fine management. Under the new situation, the extensive application of the Internet of Things, big data and the development of networked and intelligent infrastructure control systems have led to the deep integration of physical space and cyberspace, and the innovative development of artificial intelligence technology has brought new solutions for smart city management [9-11]. How to fully utilize modern information technology, apply the innovative strategies of artificial intelligence technology in smart city management, improve the level of wisdom and refinement of urban management work, and enhance the effectiveness of urban management and law enforcement has become a realistic topic in the field of urban management [12-14].

The article describes the city’s intelligent management strategy, then according to the actual situation using hierarchical analysis method for the establishment of the seventeen intelligent city management evaluation indicators for subjective assignment, the entropy method to determine the objective weights, taking into account the intelligent management of the city A data has the variability of the hierarchical analysis and the entropy method combined with the intelligent management of the city A indicators of the comprehensive assignment, to achieve a more scientific and accurate level of intelligent management. Evaluation. Finally, by applying the superiority and inferiority solution distance method, we take five different cities as evaluation objects and analyze the proximity between them and the corresponding targets for the comparative analysis of the intelligent city management level, thus realizing the scientific evaluation of the intelligent management level of A city.

2

Overview

Characterized by the rise of smart buildings and smart city management technologies, it marks the advent of the era of comprehensive application of high technology in urban construction. Literature [15] compares different smart city models implemented in Spain with a city that has not yet become a smart city to explore the factors that enable the transformation of a traditional city into a smart city, and the results of the experiment point to the fact that in order for an average medium-sized city to become a smart city, it is necessary to implement a comprehensive dashboard containing different recommendations in the strategic areas analyzed. Literature [16] in order to examine the relationship between technology-oriented knowledge management, innovation, e-government and smart city performance, quantitative analysis using knowledge management based service science theory and diffusion of innovations theory, the results show that the relationship between knowledge management, innovation, e-government and e-service delivery is not only direct, but also contextual and interactive, a finding that contributes to smart city management Sustainable Development. Literature [17] shows that the convergence of IoT technologies and AI will transform the infrastructure of smart cities, improving the sustainability, productivity and comfort of city dwellers, and emphasizes the potential contribution of 5G networks and AI in advancing the modern urban environment, and the study paves the way for the development of smart city management, which will greatly improve the quality of life of city dwellers. Literature [18] discusses the application of AI in smart city waste management, noting that it not only saves transportation distance, cost, and time in waste logistics, but also identifies and classifies waste, improves waste pyrolysis, carbon emission estimation, and energy conversion, and helps to improve the urban waste ecosystem. Literature [19] describes the components, structural support, infrastructure, and superstructure of smart city and explores the implementation of smart city and its impact on better city management with Indonesia as the object of study, aiming to provide reference value for the implementation of better smart city management strategies. Literature [20] utilized literature analysis and case studies to pinpoint the present advanced technologies for smart city management and proposed a new model of strategic smart city management that is capable of sustaining smart city development. Literature [21] proposes an end-to-end air quality prediction model for smart city applications based on machine learning techniques and deep learning techniques, and verifies its effectiveness through empirical analysis, which provides insights into the most critical factors affecting air quality, and is of great significance for air quality monitoring and management in smart city management.

3

Application of artificial intelligence technology in urban management

3.1

Application of digital technology in urban management

Optimize city management strategies by combining 5G technology. Relative to 4G technology, 5G technology is more reliable and less delayed, and its delay can be on the order of ms. And 5G technology has the characteristics of ultra-high transmission rate, ultra-low latency and ultra-large connection, which can solve the demand for large bandwidth and low latency in various application scenarios, and enable the realization of low-cost and low-latency node access technology. Utilizing different slicing technologies and edge technologies, network resources can be flexibly configured to meet the needs of different application scenarios, thus achieving the goal of flexible deployment.

Apply artificial intelligence technology to improve the efficiency of urban management. At this stage, artificial intelligence technology has become an important means of improving urban management, which is capable of predicting and making decisions on various types of behaviors, and ultimately realizing fast and precise execution and operation. Artificial Intelligence (AI) is a data-value-based operation based on Big Data, which is all about “thinking” and “decision-making”, and is inextricably linked to technologies such as cloud computing.

Comprehensive use of digital technology to build a map, a network and a screen. “One Map, One Network and One Screen” refers to the unification of data resources collection and analysis, IOT monitoring and command and dispatching based on 2D/3D digital maps, so as to provide multi-source data and scientific decision-making support for the practical application in the fields of urban security risk prevention and control, emergency command and monitoring and dispatching.

3.2

New ways in AI-enabled city management

The history of human evolution and development is a history of human beings making and using tools, and different tools represent the evolutionary level of human beings. From the Stone Age, the Iron Age, the Steam Age, the Electric Age, the present Information Age and the future Intelligent Age, we use more advanced and convenient tools to change production and life. The famous architect Liang Sicheng once said: city construction is a science, the city is an organic life form, it has meridians, pulses and texture like human body, and the city’s annual cycle will grow year by year.

In the field of urban management, the use of artificial intelligence + security, urban management, municipal, emergency and other core applications provides more effective management methods and management tools for urban management. AI + urban management analyzes and identifies people, vehicles, and objects through intelligent algorithms, conducts structured analysis of massive view data, automatically detects and captures events that violate the rules of urban management, and forms a business closed-loop of data collection, case establishment, case dispatch, case The business closed loop of data collection, case establishment, case processing, case dispatch, case dispatch and case feedback will greatly liberate the front-line law enforcement force on the road, realize the change of working mode from personnel on-site inspection to platform alarm assistance, and make the urban management of public facilities, urban sanitation, emergencies, street order and other urban management more efficient. ai + security, through the intelligent identification of pedestrians and license plates, car models, car bodies, car labels, and other vehicle information, it effectively records the passing of AI + security, through the intelligent recognition of vehicle information such as vehicle license plate, vehicle model, body, marking, etc., effectively records the vehicles and pedestrians passing by, achieving the effect of leaving license plates after passing by vehicles, and leaving pictures after passing by people. Intelligent algorithms can also intelligently identify the behavior of vehicles and pedestrians, and curb illegal signs, markings, signals, and other illegal behaviors, including vehicles running red lights, running green lights, speeding, changing lanes, and stopping, and improve the behavior of pedestrians who run red lights, don’t let pedestrians go, don’t wear seat belts, and drive and talk on the phone.

4

Evaluation modeling for smart city management

4.1

Hierarchical analysis to calculate subjective weights

Hierarchical analysis method [22] is a multi-level decision analysis method, which is based on the overall strategic objectives. First of all, the problem should be stratified according to the actual situation from the three levels of goal level, guideline level and program level, and analyzed for each level through expert distribution, and finally get the decision-related results. In this paper, the hierarchical analysis method will be used to subjectively assign weights to the indicators. First, based on the various impacts produced by the intelligent management behavior of City A, the hierarchical structure of intelligent management performance evaluation of City A is constructed. Then, the judgment matrix of indicators at each level is constructed, and the relative importance between indicators is judged by expert scoring method: finally, the consistency of the matrix is judged and the weights of indicators are determined. The process of determining subjective weights using hierarchical analysis is as follows:

1)

Construct the hierarchical structure

Constructing a hierarchical structure of a perfect system helps to solve the problem. The hierarchy exists from top to bottom in a layer-by-layer dominance relationship. In this paper, the final evaluation goal, i.e., environmental performance evaluation, is taken as the target layer, and the evaluation indexes of urban intelligent management are refined layer by layer to the guideline layer and indicator layer based on the balanced scorecard. The specific hierarchical structure is shown in Table 1.

Table 1.

The intelligent city management performance evaluation index

Target layer	Criterion layer	Index layer
A city management performance evaluation of smart city A	Infrastructure indicators B1	Fiber optic broadband access rate C1
		Wireless network coverage C2
		Urban tube intelligent terminal coverage C3
		Urban management public platform and basic database coverage C4
	Intelligent government index B2	Information disclosure timely rate C5
		Information security level C6
		Intelligent decision penetration rate C7
	Huimin service index B3	Urban management case rate C8
		Life domain service application C9
		Urban management service satisfaction C10
		Supervision and complaint rate C11
	Guarantee index B4	Policies and regulations C12
		Organization system improvement C13
		Capital input level C14
	Developmental index B5	Project overall development sustainable performance C15
		Environment and resources sustainable C16
		Service rate growth rate C17

(1)

Infrastructure indicators

Infrastructure indicators are one of the most important and fundamental indicators for performance evaluation in the current smart city management construction. For any smart city management construction, infrastructure construction is the foundation to ensure the successful completion and operation of the entire smart city management system. Therefore, infrastructure indicators are of great importance.

(2)

Smart Government Indicators

As a public construction project, the construction of smart city management represents the level of government work and government image of the entire local government, so the smart government indicators have also become an important indicator of its performance.

(3)

Indicators of Beneficial Services

As smart city management is essentially a social public project, the ultimate goal of the project’s construction is to meet the needs of the general public. Therefore, the people service indicator is also one of the important indicators in the evaluation of smart city management performance.

(4)

Guarantee Indicators

Smart city management before, during and after the construction of the delivery of the use of a large number of funds, material resources and other inputs, only in order to ensure the smooth construction and stable operation in the later stage.

(5)

Developmental Indicators

To realize the sustainable and stable development of society, it is necessary to adhere to the principle of sustainable development, and because of its own public nature, smart city management should pay more attention to sustainable development.

2)

Constructing judgment matrix

Next, the judgment matrices at the criterion level and program level are constructed to determine the relative importance of the two indicators. And by distributing the matrix questionnaire to experts in related fields, the 1-9 scale method is used to quantify the relative importance of the indicators, and then the judgment matrix is formed.

After the judgment matrix is constructed, it is necessary to normalize each column of the matrix, and then carry out row and normalization, and finally derive the maximum eigenvector and the maximum eigenroot. Among them, the maximum eigenvector can be used as the weight vector of the index.

3)

Consistency test

After the confirmation of the weights of each indicator, it is necessary to test the consistency of the matrix. Due to the existence of a certain degree of complexity of objective things, it is difficult to fully meet the consistency of the judgment matrix, if the consistency of the judgment matrix is not acceptable, then the judgment matrix needs to be readjusted until the consistency of the judgment matrix falls within the acceptable range. Among them, consistency refers to the logical consistency of the scoring results when experts score the degree of importance of the indicators. The consistency index of the judgment matrix is CI value, which is mainly used to judge the degree of tolerance of inconsistency between a certain layer of indicators and the previous layer of indicators, CI=0 indicates that the judgment matrix fully meets the consistency, and the larger the CI value is, the worse the consistency is. In order to identify the matrix consistency is reasonable, usually need to calculate the random consistency ratio CR value. The specific steps are as follows:

First, determine the consistency test index. The formula is as follows: (1) $C I = \frac{λ_{\max} - n}{n - 1}$

In the formula:

λ_max-- The largest characteristic root of the judgment matrix.

Secondly, find the average random consistency index value in the RI value table, and find the average random consistency index value RI of the same order as the judgment matrix.

Finally, calculate the stochastic consistency ratio CR, and determine whether the judgment matrix passes the consistency test according to the CR value. The formula is as follows: (2) $C R = \frac{C I}{R I}$

When CR ≤ 0.1, the consistency test is passed; when CR ≥ 0.1, it is considered that the judgment matrix weights are not logical and cannot pass the consistency test, and the judgment matrix needs to be adjusted until the consistency test is within the acceptable range.

4.2

Entropy weighting method to calculate objective weights

The calculation steps of entropy weight method [23] usually include data preprocessing, normalization of indicator values, determination of entropy of evaluation indicators, definition of entropy weights and calculation of weight values of the system. The whole calculation process is supported by objectively existing real data without any subjective human factors. Therefore, the entropy weight method can be used to determine the objective weights. The specific procedure is as follows:

1)

Normalize the original matrix

Normalizing the indicator data can eliminate the influence of different indicator scales. Assuming that there are n evaluation object and m evaluation indicators, the matrix that has been normalized is X = (x_ij)_mn, and the normalized matrix will be $X^{'} = {(x_{i j}^{'})}_{m m}$ after normalization of the original matrix, of which x_ij represents the original data of the i th evaluation indicator of the j th evaluation object. $x_{i j}^{'}$ represents the normalized data of the i th evaluation indicator of the j th evaluation object.

For positively normalized indicators, i.e., indicators that have a positive effect on the evaluation, the normalization formula is: (3) $x_{i j}^{'} = \frac{x_{i j} - \min (x_{i})}{\max (x_{i}) - \min (x_{i})}$

For negatively oriented indicators, i.e., indicators that have a negative effect on the evaluation, the standardized formula is: (4) $x_{i j}^{'} = \frac{\max (x_{i}) - x_{i j}}{\max (x_{i}) - \min (x_{i})}$

2)

Find the ratio of each indicator under each program

Calculate the share of the j nd object of the i st indicator in the total value P_ij and calculate the size of the variation of each indicator. The formula for calculation is: (5) $P_{i j} = \frac{x_{i j}^{'}}{\sum_{j = 1}^{n} x_{i j}^{'}}, i = 1, \dots, m, j = 1, \dots, n$

3)

Calculate the entropy value e_i

The entropy value of the i nd indicator is calculated by the formula: (6) $e_{i} = - \frac{1}{\ln n} \sum_{j = 1}^{n} P_{i j} \ln P_{i j}, i = 1, \dots, m, j = 1, \dots, n$

4)

Calculate the weight of each indicator

The weight of the i th indicator can be calculated by the following formula: (7) $ω_{i} = \frac{(1 - e_{i})}{\sum_{i = 1}^{n} (1 - e_{i})}$

4.3

Hierarchical analysis-entropy weighting method to calculate composite weights

The combination of hierarchical analysis method and entropy weight method for the assignment of the indicator system of intelligent management in City A takes into account both the characteristics and actual situation of intelligent management in City A and the data changes of intelligent management in City A, which can evaluate the evaluation level of intelligent management in City A more scientifically and accurately. Through reading a large amount of literature, it is found that most scholars use the multiplicative normalization method and the arithmetic average method when performing the comprehensive weight calculation. At present, the exploration of the multiplication normalization method is still weak, and its direct multiplication operation tends to further amplify the weight difference, making the original larger weights further expand and the smaller weights further shrink, which tends to easily cause the extremity of the distribution of indicator weights, and then lead to the irrational distribution of weights. Therefore, this paper will choose the arithmetic average method to determine the comprehensive weights, and the calculation formula is as follows: (8) $σ_{1} = \frac{γ_{1} + ω_{1}}{2}$

4.4

Comprehensive analysis of the TOPSIS method

The basic process of TOPSIS evaluation [24] is as follows: 1)

When encountering a multi-attribute decision problem, there are generally m evaluation objective D₁,D₂…D_n and n evaluation indicators for each objective X₁,X₂,…,X_n. The canonical decision matrix is obtained by vector programming and is set to be the X = (x_ij)_m×n canonicalized matrix R = (r_ij)_m×n. (9) $r_{i j} = \frac{x_{i j}}{\sqrt{\sum_{i = 1}^{m} x_{i j}^{2}}}, i = 1, 2, \dots, m; j = 1, 2, \dots, n$

2)

Construct the weighted data matrix.

By calculating the weight normalized value r_ij, establish the weight normalized moments with respect to the weight normalized value r_ij. (10) $v_{i j} = w_{j} \times b_{i j}, i = 1, 2, \dots, m; j = 1, 2, \dots, n$

where w_j is the weight of the first indicator.

3)

Determine the positive and negative ideal solutions.

Determine the positive ideal solution and negative ideal solution based on the weight normalization value v_ij. (11) $Positive ideal solution V^{*} = {\begin{array}{l} m a x_{i} v_{i j}, j \in J_{1} \\ m i n_{i} v_{i j}, j \in J_{2} \end{array}, i = 1, 2, \dots, m$ (12) $Negative ideal solution V^{-} = {\begin{array}{l} m i n_{i} v_{i j}, j \in J_{1} \\ m a x_{i} v_{i j}, j \in J_{2} \end{array}, i = 1, 2, \dots, m$

Where, J₁ has the benefit type attribute, indicating the optimal on the i nd index; J₂ has the cost type attribute, indicating the worst value on the i th index.

4)

Calculate the distance from each evaluation program to the positive and negative ideal solutions. The distance from the target to the positive ideal solution A^* is S^*, and the distance to the negative ideal solution A⁻ is S⁻: (13) $S^{*} = \sqrt{\sum_{j = 1}^{n} {(v_{i j} - v_{j}^{*})}^{2}}, i = 1, 2, \dots, m$ (14) $S^{-} = \sqrt{\sum_{j = 1}^{n} {(v_{i j} - v_{j}^{-})}^{2}}, i = 1, 2, \dots, m$

5)

Calculate the relative fitness of each evaluation scenario to the positive and negative ideal solutions C^*: (15) $D_{i}^{*} = \frac{s_{i}^{-}}{(s_{i}^{*} + s_{i}^{-})}, i = 1, 2, \dots, m, D^{*} \in (0 - 1)$

6)

Rank the ideal solutions according to their closeness D^* size.

Rank the values of D^* in descending order to get the ranking of each evaluation target. Ranking results The larger the value of D^*, the better the target, and the highest value is the optimal selection target.

5

Empirical Study on Evaluation of Smart Management Performance in City A

The confirmation of the weights of the selected indicators is based on the hierarchical analysis method. This study invites experts to score each of the taken hierarchical indicators two by two, and the results obtained from the scoring are processed to obtain the final scoring values. According to the numerical value substitution calculation, empirical verification, through the consistency test. And derive the stratified weights and summary weights of each indicator, the weight results are shown in Table 2, the highest proportion of the weight of the security indicator in the guideline layer is 0.3584, and the lowest proportion of the indicator of the people’s services is only 0.0879.

Table 2.

Level analysis index weight summary table

Target layer	Criterion layer	Index layer	Hierarchical weight	Weight
A city management performance evaluation of smart city (1.0000)	Infrastructure indicators (0.1354)	Fiber optic broadband access rate	0.2589	0.0525
		Wireless network coverage	0.2512	0.0524
		Urban tube intelligent terminal coverage	0.1958	0.0343
		Urban management public platform and basic database coverage	0.2941	0.0566
	Intelligent government index (0.2356)	Information disclosure timely rate	0.2458	0.0505
		Information security level	0.3957	0.0804
		Intelligent decision penetration rate	0.3585	0.0753
	Huimin service index (0.0879)	Urban management case rate	0.2854	0.0527
		Life domain service application	0.2105	0.0404
		Urban management service satisfaction	0.2465	0.0509
		Supervision and complaint rate	0.2576	0.0515
	Guarantee index (0.3584)	Policies and regulations	0.3156	0.0712
		Organization system improvement	0.4178	0.0801
		Capital input level	0.2666	0.0513
	Developmental index (0.1827)	Project overall development sustainable performance	0.3468	0.0756
		Environment and resources sustainable	0.2987	0.0501
		Service rate growth rate	0.3545	0.0715

After obtaining the indicator weights based on AHP, the indicator weights are then determined by the entropy weight method. Firstly, the data of the initial data matrix are calculated and organized to confirm the entropy value and calculate the hierarchical weight of each indicator and the weight value based on entropy weight method according to the formula. The index weight values obtained based on the entropy weight method and the index weight values obtained based on hierarchical analysis method are combined. Finally, the composite weights of different indicators are obtained, and the calculation results are shown in Table 3 below, the highest composite weight is the growth rate of service visits, with a composite weight of 0.1299, and the lowest is the supervision and complaint rate, with a composite weight of 0.0159.

Table 3.

Entropy weights index weight summary table

Target layer	Criterion layer	Index layer	Hierarchical weight	Weight	Composite weight
A city management performance evaluation of smart city (1.0000)	Infrastructure indicators (0.1354)	Fiber optic broadband access rate	0.2526	0.0352	0.0445
		Wireless network coverage	0.2517	0.0351	0.0509
		Urban tube intelligent terminal coverage	0.2515	0.0381	0.0356
		Urban management public platform and basic database coverage	0.2442	0.0276	0.0789
	Intelligent government index (0.2356)	Information disclosure timely rate	0.3486	0.0899	0.0897
		Information security level	0.3301	0.0395	0.0789
		Intelligent decision penetration rate	0.3213	0.0489	0.0589
	Huimin service index (0.0879)	Urban management case rate	0.2552	0.0319	0.1075
		Life domain service application	0.2398	0.0112	0.0985
		Urban management service satisfaction	0.2514	0.0909	0.0179
		Supervision and complaint rate	0.2536	0.0815	0.0159
	Guarantee index (0.3584)	Policies and regulations	0.3108	0.0812	0.0255
		Organization system improvement	0.3440	0.0805	0.0698
		Capital input level	0.3452	0.0713	0.0442
	Developmental index (0.1827)	Project overall development sustainable performance	0.2056	0.0856	0.0281
		Environment and resources sustainable	0.3010	0.0601	0.0253
		Service rate growth rate	0.4934	0.0915	0.1299

Next, the obtained composite weights are weighted with the normalization matrix. Then calculate the distance between the indicators of five different cities and the positive and negative ideal solutions respectively, and rank the calculation results of each indicator according to the size of the value, and the obtained results of the analysis of the infrastructure indicators, the smart government indicators, the indicators of the people’s services, the safeguarded indicators, and the developmental indicators are shown in Tables 4-Table 8 respectively. The relative proximity of the above indicators is comprehensively calculated and finally the comprehensive evaluation ranking is obtained, and the results of the comprehensive evaluation ranking are shown in Table 9.

Table 4.

Infrastructure indicators results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.1489	0.166	0.2223	2
B	0.1426	0.2669	0.2742	1
C	0.1914	0.1376	0.1758	4
D	0.2756	0.1416	0.1421	5
E	0.2291	0.1792	0.1851	3

Table 5.

Intelligence policy results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.7823	0.1885	0.0806	5
B	0.039	0.8776	0.3986	1
C	0.5789	0.3284	0.1503	3
D	0.7174	0.3169	0.1271	4
E	0.4687	0.6634	0.2439	2

Table 6.

Service indicators results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.4176	0.5902	0.1801	3
B	0.5768	0.5212	0.1457	4
C	0.0635	0.8498	0.2859	2
D	0.041	0.8692	0.2929	1
E	0.746	0.3358	0.0949	5

Table 7.

Guarantee index results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.6337	0.4579	0.1479	4
B	0.6363	0.4326	0.1431	5
C	0.1044	0.9353	0.3534	1
D	0.5841	0.6111	0.1802	2
E	0.5717	0.5621	0.1753	3

Table 8.

Developmental indicator results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.2238	0.7368	0.273	2
B	0.5967	0.4853	0.1593	4
C	0.1058	0.9239	0.3554	1
D	0.4486	0.5858	0.2012	3
E	0.9097	0.029	0.0106	5

Table 9.

Comprehensive evaluation results

City	Rational distance D⁺	Negative ideal distance D^-	After normalization	Ranking
A	0.3829	0.4678	0.1948	4
B	0.3435	0.4578	0.2024	3
C	0.3165	0.5024	0.2177	2
D	0.2028	0.5398	0.2585	1
E	0.5245	0.2914	0.1269	5

Through the final calculation results, it can be found that the comprehensive evaluation ranking of the five cities is as follows: City D>City C>City B>City A>City E. According to the analysis, it can be seen that the indicators of smart government and safeguard indicators of City A are weaker, ranked 5th and 4th respectively, and need to be adjusted. Secondly, from the perspective of infrastructure indicators, people’s services indicators and developmental indicators, City A has certain advantages and is in the middle to upper level, ranking 2nd, 3rd and 2nd respectively. these different dimensions of the impact of the indicators comprehensively and objectively reflect the strengths and weaknesses of City A. Comparison and analysis of the specific indicators can help City A to examine its own shortcomings and strengths in smart city management, and to a certain extent can help City A to find a way to manage the smart city in a more efficient way. To a certain extent, it can help City A to find the problems and improvement methods in smart city management.

6

Analysis of factors affecting smart management performance

In order to understand the spatial variability of smart management in City A, the comprehensive performance of the towns and streets of smart management in City A is taken as the dependent variable, and its influencing factors are detected by using the divergence and factor detector as a single factor. The q-value calculated by the geographic detector indicates the degree of influence of the influencing factor on the spatial variability of the comprehensive performance of smart management, and the larger the q-value, the stronger the driving force of the influence of the influencing factor on the spatial variability of the comprehensive performance of smart management. The detection results are shown in Table 10, and the specific order of the q-value of each influence factor is as follows: total industrial output value (0.657) > total population (0.582) > science, education, culture, and health facilities (0.563) > forest cover (0.506) > green space and parks (0.463) > arable land (0.462) > average elevation (0.339) > total agricultural output value (0.330) > business income (0.320)>basic farmland (0.265)>number of educated people (0.259)>water area (0.222)>per capita income in rural areas (0.214)>transportation network (0.179)>public facilities (0.112), thus it can be seen that the total value of industrial output, the total population, the facilities of science, education, culture, and health, and the forest coverage rate are the main influences leading to the spatial variability of the performance of smart management in A City Factors.

Table 10.

Industrial land performance impact factor detection results

Influencing factor	Factor symbol	Influencing factor	Q value
Natural environment	X1	Mean elevation	0.339
	X2	waters	0.222
	X3	Basic farmland	0.265
	X4	ploughing	0.462
	X5	Forest coverage	0.506
Economic development	X6	Gross agricultural output	0.33
	X7	Total industrial output	0.657
	X8	Commercial income	0.32
Social livelihood	X9	Rural income per capita	0.214
	X10	population	0.582
	X11	education	0.259
Public service	X12	Educational facilities	0.563
	X13	Public facilities	0.112
	X14	Traffic network	0.179
	X15	Green and park	0.463

Based on the detection results, in order to further analyze the role of factors influencing the performance of smart management in City A, the q-value is divided into 4 grades through the natural discontinuity grading method, and the results of the grade grading are shown in Table 11.

Table 11.

Q rating table

Grade	Q value	Importance
First class	>0.504	Dominant factor
Second class	0.341-0.504	Secondary dominance
Tertiary	0.224-0.340	Important factor
Four level	<0.224	General factor

The results of the influence factor detection ranking of the spatial variability of performance in each dimension of smart management in City A. The ranking results are shown in Table 12.

Table 12.

Intelligent management of city A performance factor detection results

Economic	Q value	Social	Q value	Ecological	Q value	Ecological	Q value
X5	0.713	X7	0.654	X5	0.314	X1	0.613
X1	0.702	X5	0.519	X3	0.302	X10	0.591
X4	0.491	X12	0.449	X1	0.269	X9	0.523
X12	0.455	X10	0.444	X15	0.225	X6	0.519
X11	0.379	X1	0.374	X8	0.198	X2	0.487
X15	0.383	X2	0.334	X7	0.184	X4	0.476
X7	0.373	X15	0.326	X9	0.18	X15	0.464
X8	0.36	X4	0.264	X14	0.143	X5	0.414
X10	0.317	X14	0.265	X2	0.137	X7	0.391
X14	0.176	X9	0.247	X12	0.131	X12	0.382
X13	0.164	X8	0.243	X10	0.12	X3	0.217
X6	0.142	X6	0.188	X6	0.12	X14	0.135
X3	0.123	X11	0.181	X4	0.099	X8	0.127
X2	0.102	X3	0.179	X11	0.075	X11	0.116
X9	0.051	X13	0.091	X13	0.06	X13	0.013

In this study, the four influencing factors with the highest q value were selected for analysis, and the top four influencing factors for economic performance were forest cover (q=0.713) > average elevation (q=0.702) > cropland (q=0.491) > science, education and recreation facilities (q=0.455), which indicated that forest cover had the most significant influence on economic performance, and average elevation had a stronger influence than cropland and science, education and recreation facilities’ stronger influence.

The first four factors of social performance are industrial output value (q=0.654) > forest cover (q=0.519) > scientific, educational and recreational facilities (q=0.449) > total population (q=0.444), which indicates that industrial output value has the most significant effect on social performance, and forest cover has a stronger effect than scientific, educational and recreational facilities and total population.

The first four factors of ecological performance are forest cover (q=0.314) > basic farmland (q=0.302) > average elevation (q=0.269) > green areas and parks (q=0.225), which indicates that forest cover has the most significant effect on ecological performance, and basic farmland has a stronger effect than average elevation and green areas and parks.

The top four influencing factors of land performance are average elevation (q=0.613) > total population (q=0.591) > rural per capita income (q=0.523) > total agricultural output value (q=0.519), which indicates that average elevation has the most significant influence on land performance, and total population has a stronger influence than rural per capita income and total agricultural output value.

The use of geographic detectors can not only identify the size of the role of a single influencing factor, but also identify the size of the interaction that exists between two influencing factors, that is, evaluating when two influencing factors interact with each other, whether the interaction enhances or attenuates the spatial variability of the interaction, in this study, the use of cross-interaction detectors for the detection of cross-interaction of influencing factors, and the type of factor interactions are shown in Table 13.

Table 13.

Factor interaction type

Graphic representation	Criterion	Interaction
--†----------*----q	q(X_i ∩ X_j) < min[q(X_i),q(X_j)]	Nonlinear attenuation
------†------*----q	min[q(X_i),q(X_j)] < q(X_i ∩ X_j) < max[q(X_i),q(X_j)]	Single factor nonlinear attenuation
----------†--*----q	max[q(X_i),q(X_j)] < q(X_i ∩ X_j)	Double factor enhancement
------------†*----q	q(X_i ∩ X_j) = q(X_i) + q(X_j)	Independence
------------*--†--q	q(X_i ∩ X_j) > q(X_i) + q(X_j)	Nonlinear enhancement
*	min[q(X_i),q(X_j)]
*	max[q(X_i),q(X_j)]
*	q(X_i) + q(X_j)
†	q(X_i ∩ X_j)

Tables 14 and 15 display the results of factor interaction detection and nonlinear enhancement detection for smart management in City A, which are achieved by using geo-detectors for detection. The results show that there are numerous interactions with enhanced relationships between the various influencing factors, and there are no independent or weakened relationships.

Table 14.

Industrial land shadow factor interaction detection results

	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11	X12	X13	X14	X15
X1	0.341
X2	0.9	0.221
X3	0.873	0.445	0.266
X4	0.888	0.553	0.558	0.466
X5	0.701	0.891	0.927	0.851	0.503
X6	0.822	0.417	0.386	0.584	0.892	0.331
X7	0.871	0.83	0.796	0.942	0.899	0.811	0.664
X8	0.582	0.795	0.783	0.942	0.597	0.802	0.788	0.319
X9	0.722	0.98	0.897	0.955	0.881	0.519	0.868	0.878	0.203
X10	0.751	0.825	0.776	0.83	0.784	0.835	0.786	0.719	0.892	0.575
X11	0.792	0.902	0.846	0.831	0.86	0.592	0.836	0.919	0.524	0.858	0.263
X12	0.789	0.834	0.847	0.805	0.715	0.799	0.714	0.75	0.849	0.789	0.84	0.561
X13	0.818	0.752	0.677	0.899	0.987	0.683	0.854	0.665	0.557	0.893	0.452	0.839	0.112
X14	0.903	0.733	0.866	0.84	0.89	0.702	0.767	0.818	0.705	0.793	0.617	0.828	0.787	0.172
X15	0.748	0.809	0.715	0.838	0.713	0.598	0.804	0.645	0.711	0.78	0.666	0.651	0.618	0.789	0.462

Table 15.

Factor interaction nonlinear enhancement detection results

Factor interaction q comparison	Factor interaction q comparison	Factor interaction q comparison
q(X1 ∩ X2)>q(X1+X2)	q(X3 ∩ X7)>q(X3+X7)	q(X6 ∩ X14)>q(X6+14)
q(X1 ∩ X3)>q(X1+X3)	q(X3 ∩ X8)>q(X3+X8)	q(X7 ∩ X13)>q(X7+13)
q(X1 ∩ X4)>q(X1+X4)	q(X3 ∩ X9)>q(X3+X9)	q(X8 ∩ X9)>q(X8+X9)
q(X1 ∩ X6)>q(X1+X6)	q(X3 ∩ X11)>q(X3+11)	q(X8 ∩ X11)>q(X8+11)
q(X1 ∩ X9)>q(X1+X9)	q(X3 ∩ X13)>q(X3+13)	q(X8 ∩ X13)>q(X8+13)
q(X1 ∩ X11)>q(X1+11)	q(X3 ∩ X14)>q(X3+14)	q(X8 ∩ X14)>q(X8+14)
q(X1 ∩ X13)>q(X1+13)	q(X4 ∩ X8)>q(X4+X8)	q(X9 ∩ X10)>q(X9+10)
q(X1 ∩ X14)>q(X1+14)	q(X4 ∩ X9)>q(X4+X9)	q(X9 ∩ X12)>q(X9+12)
q(X2 ∩ X5)>q(X2+X5)	q(X4 ∩ X11)>q(X4+11)	q(X9 ∩ X13)>q(X9+13)
q(X2 ∩ X8)>q(X2+X8)	q(X4 ∩ X13)>q(X4+13)	q(X9 ∩ X14)>q(X9+14)
q(X2 ∩ X9)>q(X2+X9)	q(X4 ∩ X14)>q(X4+14)	q(X10 ∩ X13)>q(X10+13)
q(X2 ∩ X11)>q(X2+11)	q(X5 ∩ X9)>q(X5+X9)	q(X11 ∩ X14)>q(X11+14)
q(X2 ∩ X13)>q(X2+13)	q(X5 ∩ X11)>q(X5+11)	q(X12 ∩ X13)>q(X12+13)
q(X2 ∩ X14)>q(X2+14)	q(X5 ∩ X13)>q(X5+13)	q(X12 ∩ X14)>q(X12+14)
q(X2 ∩ X15)>q(X2+15)	q(X5 ∩ X14)>q(X5+14)	q(X13 ∩ X14)>q(X13+14)
q(X3 ∩ X5)>q(X3+X5)	q(X6 ∩ X8)>q(X6+X8)	q(X14 ∩ X15)>q(X14+15)
q(X3 ∩ X6)>q(X3+X6)	q(X6 ∩ X13)>q(X6+13)

For example, the two factors of public facilities X13 and transportation network X14 have a q-value of only 0.112 and 0.172 alone, but when the two factors interact together, the q-value is 0.789, showing the influence effect of “1+1>2”, indicating the spatial difference of public facilities and transportation network enhancement on the intelligent management performance of city A. Therefore, the joint action of multiple factors can enhance the occurrence of spatial differentiation of smart management performance in City A, and the two-by-two interaction is much larger than the effect of single factor. In addition, the four factors of commercial income X8, rural per capita income X9, public facilities X13 and transportation network X14 intersect with multiple other factors to appear q value nonlinear enhancement, indicating that the effect of these factors alone is not obvious, and the joint effect with other factors is very active. For example, transportation network X14 alone does not significantly affect smart management in City A. The enhancement is significant when linking business services, rural production, education, public facilities, green spaces, and parks.

7

Conclusion

This study establishes a complete and effective performance evaluation system for smart city management, and discusses the rationality and effectiveness of the application of artificial intelligence technology in city management.

The highest proportion of weight in the criterion layer is the safeguard index and the lowest is the beneficiary service index. The highest and lowest composite weight shares are the growth rate of service visits and the supervision and complaint rate, which have composite weights of 0.1299 and 0.0159, respectively.

The composite evaluation rankings of the smart management performance of the five selected cities are: city D>city C>city B>city A>city E. City A’s smart government indicators and safeguard indicators are weaker, and there is room for improvement. It has advantages in infrastructure indicators, beneficiary service indicators and developmental indicators, and is in the middle to upper level, ranked 2nd, 3rd, and 2nd respectively. These data comprehensively and objectively reflect the strengths and weaknesses of City A, and help City A find the problems in smart management and further improve them.

The results of single-factor probing conclude that the total industrial output value, total population, science, education, culture and health facilities, and forest coverage are the main influencing factors for the spatial variability of smart management performance in City A. The factors of public facilities and transportation networks act alone as the main influencing factors. The q-value of the public facilities and transportation network factors acting alone is less than 0.2, but when the two factors act crosswise, the q-value reaches 0.789, indicating that there is a spatial variability of public facilities and transportation network enhancement on the smart management performance of City A. The results also demonstrate that there is spatial variability in the performance of smart management in City A, due to multiple factors acting together. It also proves that the combined effect of multiple factors is much larger than the effect of a single factor in enhancing the occurrence of spatial differentiation of smart management performance in City A.

Idioma:: Inglés

Calendario de la edición:: 1 veces al año
Temas de la revista:: Ciencias de la vida, Ciencias de la vida, otros, Matemáticas, Matemáticas aplicadas, Matemáticas generales, Física, Física, otros

RSS Feed de revista

Research on Innovative Strategies of Artificial Intelligence Technology in Smart City Management and Its Practical Effects

Zhisong Wu

Publicado en línea: 19 mar 2025

Recibido: 26 oct 2024

Aceptado: 12 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0546

Palabras claveSmart city management, Hierarchical analysis method, Entropy weight method, Topsis model

© 2025 Zhisong Wu, published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Palabras clave
Smart city management, Hierarchical analysis method, Entropy weight method, Topsis model