Research on digital sharing mechanism of agricultural science and technology resources considering collaborative innovation capability

: The digital sharing of agricultural scientific and technological resources is the key to promoting agricultural scientific and technological innovation. Explore the strategic mechanism for the digital sharing of agricultural scientific and technological resources, reveal the law of their behavioral evolution, and provide theoretical guidance for the practice of digital sharing of agricultural science and technology resources. Introducing the perspective theory of perceived utility to replace expected utility, constructing a stochastic evolutionary game model for the sharing parties and demand parties of agricultural science and technology resources, analyzing the dynamic evolution rules of both sides of resource sharing, and analyzing the key factors affecting the stability of agricultural science and technology resource sharing through numerical simulation.


Introduction
The integration of digital technology with agriculture and the countryside is a necessary step in the modernization of agriculture. Advances in technology can effectively contribute to economic growth. Through the sharing of agricultural science and technology resources, collaborative innovation among the various participating actors to materialize resources and produce beneficial effects has become the main way to promote agricultural development. Weak technical research, too low level of agricultural digitalization; lagging behind in technology promotion, lack of recognition by agriculture-related enterprises and farmers, etc. making the development of key digital agriculture technologies still face a lot of bottlenecks [1], disconnect between supply and demand of agricultural resources as a barrier to rural economic development [2]. The emergence of these problems is ultimately due to the mismatch between traditional production methods and the development of agricultural modernization, and the difficulty of meeting the needs of modern agricultural development in terms of the supporting power of science and technology and the way they are supported. It also shows that the modernization of agriculture is even more dependent on advances in agricultural science and technology. Agricultural science and technology innovation not only needs to consider its innovation capabilities, but also requires collaborative innovation across organizations, departments, industries, and regions. Through communication, coordination, cooperation, and synergy, collaborative innovation in agricultural science and technology integrates various types of agricultural science and technology resources to break down the barriers between relevant interest subjects and maximize the benefits of each subject through the effective sharing of resources. The sharing of agricultural science and technology resources is the process of transferring resources through the cooperation of various innovative entities, combined with complementary resources, to absorb and transform agricultural science and technology resources and apply them to agriculture. When it comes to collaborative innovation in agricultural science and technology, due to the complex environment of resource sharing, a large number of participants in collaborative innovation, and the uncertainty of innovation results, collaborative innovation subjects will make decisions based on their interests in the process of resource integration and consolidation. Therefore, studying the behavioral decisions of stakeholders related to the digital sharing of agricultural science and technology resources and exploring the dynamic evolution of their behavioral strategies can effectively raise the awareness of agricultural science and technology resource sharing among stakeholders, stimulate the synergistic innovation effect, and thus provide a theoretical basis for promoting agricultural science and technology progress.

Literature Review
Some scholars have conducted research on the issue of "agriculture". In the area of collaborative innovation in agriculture, Yu et al [3] proposed that the best model for collaborative innovation in agricultural enterprises is to build equity-based collaborative R&D institutions with universities and research institutes. Wang et al [4] concluded that the development of agricultural science and technology innovation in China exhibits obvious spatial correlation and spillover effects, in which differences in geographical proximity, differences in government support, and differences in agricultural economic development affect the spatial correlation of agricultural science and technology. Yang et al [5] measured the human and financial resource mismatch index in different regions and industries to explore the impact of agricultural financial resources and agricultural human resource mismatch on the output of science and technology. Sheng et al [6] argued that the use of blockchain can achieve credibility, collaboration, and interconnection of organizations in the development of agricultural industrialization, making the supply chain available with high-quality data. And based on the integration and innovation of blockchain multitechnology, can improve the efficiency of agricultural production and promote the development of collaborative innovation in agriculture. In terms of the agricultural science and technology resource sharing model, Zhu et al [7] analyzed agricultural data sharing technology and established a big data sharing model combined with blockchain technology to improve the problem of regional and independent information of major agricultural information websites significantly. Zhao [8] applied cloud computing to agricultural resource sharing, making the sharing of agricultural science and technology resources supported by a more complete information technology foundation and more extensive interconnection so that recipients of science and technology resources can enjoy more intimate services. From the perspective of costs and benefits, Zhang et al [9] argued that intelligent production in agriculture can reduce costs, reduce damage to the environment, and enable more effective use of existing scientific and technological resources. The evolutionary game is suitable for balancing the interests of various stakeholders in the sharing process of agricultural scientific and technological resources to find the best way to share scientific and technological resources. In terms of the application of evolutionary games, Wan et al [10] constructed and analyzed a conflict evolution game model, believing that it is possible to maintain a moderate state of control by controlling constructive conflicts, establishing a mutual trust communication environment and convenient communication channels in scientific and technological innovation teams, and cooperate with scientific and technological innovation teams to improve their collaborative innovation capabilities. Liu et al [11] established a game model between agricultural enterprises, the government, and farmers. Through simulation, it is concluded that reasonable government subsidies and carbon taxes for agricultural enterprises and farmers can improve the enthusiasm of agricultural enterprises and farmers to participate in low-carbon agriculture. Through reviewing existing research, it is found that although some literature has combined evolutionary game theory to study the sharing and collaborative innovation behavior of agricultural science and technology resources, there is a lack of understanding of the impact of risk aversion, risk preference, and loss sensitivity parameters on the decision-making of each sharing entity. At the same time, few scholars have considered the uncertainty of the digital environment for agricultural science and technology resource sharing. Because of this, this study introduces Gaussian white noise to characterize various random disturbances in uncertain environments, and combines prospect theory with evolutionary game theory to consider both the rational shortcomings and preferences of decision-makers. It explores resourcesharing behavior strategies among different actors in the process of agricultural science and technology resource sharing, examines the evolution rules of agricultural science and technology resource digital sharing systems, and uses MATLAB simulation to explore the main impact parameters, Based on the evolutionary mechanism of multi-agent resource sharing, this paper explores the management implications for promoting agricultural science and technology resource sharing.

Description and definition of the problem
The sharing parties of agricultural science and technology resources are mainly agriculture-related enterprises, agricultural research institutes, agricultural universities, and governments. They are mainly responsible for the research and development of new technologies, new knowledge, and new varieties. The recipients of agricultural science and technology resources are mainly organizations or individuals involved in agricultural production, including large agricultural households, ordinary farmers, agricultural cooperation institutions, and small and medium-sized agricultural enterprises.
(1) In the process of digital sharing of agricultural science and technology resources, assuming that the probability of agricultural science and technology resource sharing subjects choosing to share is x , the probability of not sharing is 1 x  , and the probability of agricultural science and technology resource receiving subjects choosing to cooperate is y , and the probability of choosing not to cooperate is 1 y  , and [0,1] The perceived utility of the prospect theory is replaced by the expected utility, and the objective probability is replaced by the decision weight function to make it more realistic.
denotes the subjectively perceived value of the difference between the actual benefit i x and the benefit 0 x at the reference point by the digitally shared interest subjects of agricultural science and technology resources after the event occurs.
 denotes the risk preference coefficient,  denotes the risk aversion coefficient, and the smaller the values of  and  , the more sensitive the attitude toward risk is,    denotes the loss aversion coefficient, and the larger its value indicates the higher sensitivity to loss of the relevant interest subject. Taking (0,0) as the reference point, the strategy choice of the sharing party and the receiving party of agricultural science and technology resources is (no sharing, no participation), and their perceived value is zero.
(3) The sharing and receiving subjects can transform agricultural science and technology resources into collaborative innovation benefits through resource sharing and cooperation, and the output of their benefits is jointly determined by the shared stock of science and technology resources Q and the coefficient of Assuming that the proportion coefficient of collaborative revenue of the resource-sharing party is (0 the collaborative innovation revenue can be expressed as fQ  , and the revenue of the resource recipient can be expressed as (1 ) f Q   . Meanwhile, resource sharing parties will invest corresponding financial, human and resource costs in the process of agricultural science and technology resource sharing, denoted by 1 C , and resource receiving parties will also incur costs in conducting resource collection and storage, denoted by 2 C .
(4) Due to the complexity of agricultural science and technology resources in the process of digital sharing and the dynamic turnover among game subjects, with 3 C denoting the loss arising from resource spillover and 1 (0 1) p p   . (5) and denote the government's financial support to the resource-sharing party and the resource-receiving party, respectively. Based on the above analysis, the prospect matrix of digital sharing of agricultural science and technology resources is shown in Table 1 Resource Recipients:

Setup of stochastic evolutionary game model
With the help of stochastic analysis theory, a stochastic interference term is introduced to portray the interference caused by uncertainties on the digital sharing of agricultural science and technology resources. 1 1 1 1 2 Equations (11) and (12) are one-dimensional stochastic differential equations containing Gaussian random disturbance terms, which are more difficult to solve directly, but the existence and uniqueness of the solutions of the differential equations can be obtained under specific conditions.

Existence and Stability Analysis of Equilibrium Solutions
. The stability discriminant for equations (11) and (12) is based on the following: denotes the zero solution moment exponential stability of the stochastic evolution equation for the shared agricultural science and technology resources. 11 12 it means that the zero solution moment index of the stochastic evolution equation for the shared side of agricultural science and technology resources is unstable.
then it means that the zero solution moment index of the stochastic evolution equation for the shared side of agricultural science and technology resources is unstable.

Simulation Analysis
Using the Milstein higher-order method, the equations are discretized as follows: Initial value setting. Assume that the initial ratio of resource sharing group and resource receiving group is 0 0 x y  =0.5 The digital sharing volume of agricultural science and technology resources is Q =550, the coefficient of collaborative innovation capacity  =0.6 The ratio of collaborative benefit distribution is f =0.6; the subsidies given by the government to resource sharing party and receiving party are 1 ; the long-term The probability of gain or loss 1 p = 2 p =0.5; the perception coefficient  =0.5.
(1) The effect of random interference intensity on the evolution process Figure 1 The effect of random disturbance intensity on the evolutionary results From Figure 1, it can be seen that the evolution of the sharer to 0 will be faster compared to the resource recipient because the sharer has all kinds of agricultural science and technology resources, and it is in a dominant position, it can make decisions that maximize its benefits more quickly. Meanwhile, the resource recipient is in a subordinate position in the game, which will slow down the evolution of the resource recipient to a steady state due to information asymmetry. Under the influence of random disturbances, the sharing and receiving parties show fluctuations of a certain magnitude in the process of evolving to a steady state. From Figs. 2(a) and 2(b), it can be seen that the evolution of the "sharing" strategy in the direction of resource sharing gradually slows down as the intensity of random interference increases. This is because the evolution to the steady state is a fluctuating state due to the uncertainty, which limits the decision-making of the sharing party to a greater extent, and thus leads to a slower evolution to the steady state. For the recipient of the resource, environmental uncertainty accelerates its convergence to zero due to its subordinate status.
(2) The impact of collaborative innovation capability on the evolution process As can be seen from Figure 2, with the increase of the coefficient of collaborative innovation capability, the evolutionary results of resource-sharing parties and resource-receiving parties gradually converge from (0,0) to (1,1). exists a critical value located at 0.6-0.7, and when is greater than this critical value, resource-sharing parties and receiving parties are more inclined to choose (sharing, cooperation) strategy, and the larger the value of , the faster the convergence speed. Therefore, the stronger the collaborative innovation ability among subjects, the higher the collaborative innovation gain will be, and the more favorable the agricultural science and technology resource-sharing behavior will be.
x(t) y(t) Figure 2 The effect of co-innovation capacity coefficient δ on the evolutionary results (3) The Impact of loss-sensitive coefficient on the evolutionary process According to prospect theory, game players are more sensitive to lose coefficients than profit coefficients, 1。 As can be seen from Figure 3, with Gradually increasing, and y gradually evolve from converging to 1 to converging to 0, and there is a critical value between 2.0 and 2.5, if Below this threshold, and converge to 1, and as The convergence speed will also increase with the decrease of. Therefore, the smaller the loss sensitivity coefficient of the resource-sharing party, the easier it is to choose the "sharing" strategy. Figure 3 The effect of loss sensitivity coefficient on the evolutionary results (4) The Impact of financial support on evolutionary processes As can be seen from Figure 4, with the increase in government funding support, resource-sharing parties, and resource recipients have gradually evolved from a selection (non-sharing, non-cooperation) strategy to a (sharing, cooperation) strategy. And there are two threshold values located at 25-30 and 10-15, respectively. When the government's financial support to the resourcesharing party and the resource recipient is greater than these two threshold values, both are more inclined to choose (sharing, cooperation) strategies and when these two threshold values are exceeded, the greater the government's financial support, the faster the convergence speed of the two will be. Therefore, increasing government funding support can promote the enthusiasm for digital sharing and cooperation of agricultural science and technology resources among game players. Figure 4 The effect of government subsidies on evolutionary outcomes x(t) y(t)

Conclusions and Recommendations
Considering the long-term nature and uncertainty of digital sharing of agricultural science and technology resources, a stochastic evolutionary game model of agricultural science and technology resources sharing parties and receiving parties is constructed, and Gaussian white noise is introduced to portray various types of stochastic disturbances under uncertain environment, which can better reflect the uncertainty characteristics of the objective world and is more in line with the reality. The main conclusions are as follows: (1) In the process of digital sharing of agricultural science and technology resources, the sharing party always evolves to a stable strategy faster than the receiving party.
In an uncertain environment, random disturbances affect the speed of evolution to a stable strategy for both the resource sharer and the receiver, and the degree of random disturbances increases and the sharer converges to 0 more slowly. For the resource receiver, due to its subordinate position, it will converge to 0 faster due to the environmental uncertainty factors.
(2) The probability of a collaborative innovation system evolving into (sharing, cooperation) is positively correlated with the coefficient of collaborative innovation capacity and government financial support. Improving the collaborative innovation capacity of both sides of the game and increasing government financial support for networked sharing of agricultural science and technology resources can effectively promote the evolution of a collaborative innovation system into (sharing, cooperation).
The relevant management implications are as follows: (1) Improve collaborative innovation capability to increase innovation benefits. By fostering the learning ability and resource collection and integration ability to participate, members, forming a sense of collaborative innovation and establishing an adaptive talent training mechanism, an upstream and downstream chain of scientific research with knowledge and value as its core is formed. The public and private sectors are encouraged to jointly solve major problems that constrain agricultural development and give full play to the heterogeneous advantages of agricultural science and technology resources. Corresponding incentives can be used to increase members' willingness to participate and solve problems with deep uncertainty in the process of agricultural science and technology innovation. To better utilize the advantages among members and improve the output of collaborative innovation, the selection of partners needs to pay more attention to their innovation capabilities. Considering the heterogeneity of their agricultural science and technology resources, otherwise, it is difficult to achieve effective integration of resources and solve relevant technical problems through collaborative innovation.
(2) Promote the construction of rural infrastructure and strengthen the construction of science and technology infrastructure capacity. Through government-led, market-led, and social participation, use emerging technologies such as the Internet of Things, 5G, and big data to reduce sharing and receiving costs, build a digital sharing platform for agricultural science and technology resources, and promote the adoption of digital application scenarios by agricultural business entities. For the development, collection and utilization of agricultural science and technology resources, the uncertainty of the digital environment and the complexity of the data need to be considered. Therefore, it is especially important to build a complete system of basic digital resources, coordinate the construction of science and technology innovation centers and regional science and technology innovation centers, reduce the cost of digital sharing and collection of agricultural science and technology resources, and improve the efficiency of using agricultural science and technology resources.
(3) Develop a diversified guidance system and improve the agricultural support and protection system. Based on the existing policy subsidies, increase the subsidies for digital sharing of agricultural science and technology resources, reduce the cost of agricultural science and technology resource sharing between the sharing and receiving parties, expand the scope of benefits for innovation subjects, increase the forward and indirect benefits of resource sharing cooperation, and activate the momentum of agricultural science and technology resource sharing.