Bifurcation Control of a Delayed Fractional Mosaic Disease Model for Jatropha curcas with Farming Awareness

In this paper, the bifurcation control of a fractional-order mosaic virus infection model for Jatropha curcas with farming awareness and an execution delay is investigated. By analyzing the associated characteristic equation, Hopf bifurcation induced by the execution delay is studied for the uncontrolled system.(en, a time-delayed controller is introduced to control the occurrence of Hopf bifurcation. Our study implies that bifurcation dynamics is significantly affected by the change of the fractional order, the feedback gain and the extended feedback delay provided that the other parameters are fixed. A series of numerical simulations is performed, which not only verifies our theoretical results but also reveals some specific features. Numerically, we find that the Hopf bifurcation gradually occurs in advance with the increase of the fractional order, and there exist extreme points for the feedback gain and the extended feedback delay which can minimize the bifurcation value.


Introduction
Jatropha curcas is a plant that is commonly seen in tropical and subtropical areas. It can grow well in marginal and poor soil, grows fast, and lives relatively long. Meanwhile, it produces seeds with an oil content of about 27% to 40%. e oil burns with clear smoke-free flame and can be used to produce biodiesel fuel [1]. erefore, as one of the most suitable alternative renewable energy resources, it is widely planted in many countries [2,3]. However, many studies showed that Jatropha plants are hosts of mosaic virus which is a main cause of the occurrence of viruses in Jatropha curcas [4,5]. Early reports about virus infections of Jatropha plants indicated that the occurrence of Cassava mosaic virus in India causes a high disease incidence from 25 to 47%. is constitutes a major obstacle to the large scale planting of Jatropha [6]. When the Jatropha plant is infected with mosaic virus, its fruit will be attacked and the yield and the quality of oil may be severely affected.
Awareness of the disease is really necessary for farmers who plant Jatropha curcas because they can take action to prevent or mitigate the problem in time. Effective knowledge can provide proper control measures when the virus infection outbreaks. It is advisable to educate the Jatropha curcas growers with agricultural information about agronomic practices and plant protection measures [7]. Common platforms for mass media such as newspapers, magazines, radio, television, and the Internet are main ways to propagandize correct and relevant information about this crop and its disease. ere are also new technologies adopted in agricultural awareness programs [8,9]. ese are very important for the grower who is engaged in the crop production. For example, Yang et al. exploited the levels of knowledge and awareness of the side effects of insecticides to the environment and human health in regions with different farming modes and found that farmers can improve their agronomic practice to reduce environmental hazard and protect human health after properly raising awareness [10]. Le Bellec et al. studied the collaboration mechanism of growers, researchers, and other factors which can alleviate the problem we face in crop management [11]. Basir et al. proposed a mathematical model to investigate the impact of awareness programs on the protection of the Jatropha curcas plant against possible epidemic spread of the mosaic disease [3].
Generally speaking, when Jatropha curcas growers realize the harm of virus infection, they need to take some time before they can take appropriate measures to prevent or reduce the occurrence of this disease. So, it is more plausible to consider an execution time delay when building such kind of virus infection model. Basir et al. analyzed how execution time delay in accepting the awareness campaign can affect the ultimate dynamics of the mosaic disease [2], while the authors in [3] investigated the Hopf bifurcation phenomenon of an epidemic model with awareness programs due to the execution time delay. Basir et al. [12] proposed a mathematical model and analyzed the effect of awareness programs on the control of pest in agricultural practice. However, these studies are all aimed at integer-order models. In this work, we further analyze the stability and Hopf bifurcation of a fractional-order virus infection model with execution delay of awareness programs.
In recent decades, fractional calculus theory has been widely applied in biology, optical and thermal system, materials science, electromagnetic field theory, mechanical mechanics, and so on [13,14]. It has been found that fractional calculus can accurately describe the rules and development process of some phenomena in natural science. Furthermore, it is found that fractional-order differential system has the advantages of simple modeling, clear parameter meaning, and accurate description for some materials and processes with memory and genetic characteristics [15][16][17]. Hence, fractional calculus has become a new mathematical tool favored by researchers, and more and more study of practical problems introduces the theory of fractional calculus and remarkable achievements have been made [18][19][20][21][22][23][24][25][26][27][28][29][30].
For a given nonlinear system, a controller aiming at modifying the bifurcation behavior is usually designed, so as to achieve some desirable dynamical behaviors, which is often called bifurcation control [31,32]. According to the different objectives of bifurcation control, different control strategies are formulated, such as PD control, time-delayed feedback control, and hybrid control. For example, Xiao et al. [22] adopted a PD control method to control Hopf bifurcations in delayed fractional-order small-world networks, while Lu et al. [32] analysed the stability and bifurcation of a fractional-order single-gene regulatory model under a PD control law. Xu et al. [28] applied two timedelayed feedback controllers to regulate a fractional-order chaotic Genesio-Tesi model, and Huang et al. [33] designed a hybrid controller for the first time to control the Hopf bifurcation of a network model.
As an effective control scheme, time-delayed feedback control was proposed by constructing a control force using the difference between the current state and its delay value, i.e., x(t) − x(t − δ), and once the system reaches a stable state, the control force vanishes [34]. e main advantage of this scheme is that it is noninvasive and it does not require a reference system because the control is generated from the information of the system itself. Besides, this scheme can be easily implemented in the actual system. Hence, in recent years, the bifurcation controls of many dynamic systems in many fields are studied by introducing time-delayed feedback controllers.
In this paper, delayed fractional mosaic disease models for Jatropha curcas with farming awareness are proposed.
We will study the stability and bifurcation of the system and discuss the bifurcation control by introducing a timedelayed feedback controller. is paper is organized as follows. In Section 2, some basic materials regarding fractional calculus are presented. e delayed fractional mosaic disease models with and without control are proposed in Section 3. In Section 4, detailed analysis of bifurcation phenomena for the two systems is carried out. Series of numerical simulations are carried out in Section 5, which not only confirm the theoretical results we have obtained but also are complementary to those results with specific features. e paper ends with a brief conclusion.

Preliminaries
In this section, we will briefly introduce some notations and definitions about fractional calculus theory, which will be useful in the following discussion. ere are several kinds of definitions of fractional derivatives proposed in previous research, such as Riemann-Liouville fractional derivative, Grünwald-Letnikov fractional derivative, Caputo fractional derivative, Weyl fractional derivative, and Marchaud fractional derivative [35]. It is important to note that Caputo fractional derivative has the advantage of requiring the initial conditions to be easily derived from the controlled system, which make it easier to apply to practical problems [32]. erefore, in this paper, we only study Caputo fractional derivative.
Definition 1 (see [36]). e Caputo fractional-order derivative with fractional order α for a continuous function p(t): R + ⟶ R n is defined by where 0 ≤ m − 1 ≤ α < m, m ∈ Z + and Γ(·) is the Gamma function. e constant α is the value of the fractional order.

Remark 1.
Our work is based on the Caputo derivative, and for convenience, in this work, we denote the Caputo fractional-order derivative operator C D α t 0 ,t p(t) by D α p(t) and suppose that α ∈ [0, 1].

Model Formulation
e authors in [2] studied an integral order model of virus infection for Jatropha curcas with farming awareness. An execution delay is incorporated, which measures from the moment that the relevant information is available to the moment that farmers take action with these knowledge. ey mainly explored the impact of the time delay on the dynamical behaviors.
Motivated by the work in [2], in this paper, we consider the virus infection model with fractional-order and the execution delay; then, we propose the following delayed fractional-order virus infectious model: e interpretations of the variables and parameters are listed in Table 1.
To control the bifurcation caused by the execution delay, we introduce a time-delayed feedback controller as follows: en, we gain the following controlled system: Remark 2. k < 0 is the negative feedback gain and δ > 0 stands for the delay of feedback control. In the field of ecological control, with the aim of enhancing the stability performance, the farmers may harvest or transplant some plants on the basis of past data (P S (t − δ)).
We will study the existence of Hopf bifurcation and explore the impact of time delay, the fractional order, the feedback gain, and the feedback delay on the occurrence of the bifurcation.

Main Results
In this section, we firstly study the stability of the coexistence equilibrium and the Hopf bifurcation caused by the time delay for the uncontrolled system (3). en, the impact of the fractional order and the feedback control on the occurrence of the bifurcation is investigated for the controlled system (5). e system (3) has three equilibria: the plant-vector-free equilibrium E 1 � (0, 0, 0, (c/η)), the disease-free equilibrium E 2 � (K, 0, 0, (c/η)), and the coexistence equilibrium E * � (P * S , P * I , V * , M * ). By direct calculation, we can easily obtain and M * is the positive root of the following equation: where a 1 � βrnd 2 > 0, Obviously, if a 4 < 0, the above cubic polynomial equation has a unique positive root which is denoted by x � M * . en, we give the following result.

Proposition 1. Assume that
Then, system (3) has a unique coexistence equilibrium For convenience, let and these denotations will be used in the following discussion.

Bifurcation Analysis of the Uncontrolled System (3).
In this section, by choosing the execution delay τ as a bifurcation parameter, we analyse the existence of Hopf bifurcation and the critical value of the time delay at which a Hopf bifurcation occurs is obtained. By linearizing (3) at E * , we get the following Jacobian matrix: us, we have the following characteristic equation: rough simplification, we can obtain where Suppose that τ � 0, then equation (13) becomes Assume that en, by the Routh-Hurwitz criterion, we have the following result.

Proposition 2.
For τ � 0, if assumptions (H 1 ) and (H 2 ) are satisfied, then the roots of (15) are real and negative. erefore, the coexistence equilibrium E * is locally asymptotically stable.

Bifurcation Control of System
and the corresponding characteristic equation: where Assume that s � ωi(ω > 0) is a root of (11); then, we obtain 6 Complexity Separating the real and imaginary parts, we obtain where and φ i , i � 1, 2, . . . , 13 are defined by Appendix B. It follows from (40) that Similarly, we can obtain Assume that equation (40) has at least one positive real root ω 1 ; then, it follows from the second equation of (39) that Define To derive conditions for the existence of Hopf bifurcation, we make the following hypothesis: where Φ 1 , Φ 2 , Ψ 1 , and Ψ 2 are defined by equation (47).

Complexity
e proof is completed.

Numerical Simulation
In this section, two numerical examples are provided to verify our theoretical results of systems (3) and (5).
In addition, we explore the impact of the fractional order α on the bifurcation value τ 0 for the uncontrolled system (3).

Complexity
From Figure 3, we can see that the bifurcation value τ 0 is very sensitive to the change of the fractional order α. With the increase of the order, the bifurcation value decreases rapidly, which implies that the stability region of system (3) becomes smaller. In particular, compared to the corresponding integral order, the fractional system has a larger stability region.
Besides, by selecting τ � 15, we investigate the convergence rate of the system. From Figure 4, we find that, with the increase of the order α, the convergence rate of system (3) becomes slower.  Table 2, while the feedback gain and the feedback delay are selected as k � − 2 and δ � 4.8, respectively. When the order is chosen as α � 0.95, we can obtain that ω 1 � 0.0213 and the corresponding critical value τ 0 δ � 32.8007. According to eorem 2, the coexistence equilibrium E * is asymptotically stable when τ � 28 < τ 0 δ (see Figure 5), while it is unstable when τ � 33 > τ 0 δ and a Hopf bifurcation occurs (see Figure 6).

Complexity
We also explore the impact of the fractional order α on the bifurcation value τ 0 δ for the controlled system (5). From Figure 7, we can see that the bifurcation value decreases rapidly with the increase of the order, which means that the stability region of system (5) also becomes smaller. Similarly, with the rise of the order α, the convergence rate of system (5) slows down (see Figure 8).
Furthermore, we investigate the impact of the feedback gain k and the feedback delay δ on the bifurcation value τ 0 δ . By fixing α � 0.95, δ � 4.8, and τ � 30, we find that there is a  Complexity 13 minimum point around k � − 2 (see Figure 9). Specifically, with the increase of the feedback gain k, the stability region of system (5) first slowly reduces to a minimum, then gradually increases to a high level. ree representative values k � − 0.5, k � − 2, and k � − 6 are chosen in Figure 10, and we can see that system (5) displays the slowest convergence rate when k � − 2 (which is the nearest to the minimum point).
In addition, by fixing α � 0.95, k � − 2, and τ � 30, we can see from Figure 11 that there is also a minimum point around δ � 3.5. Specifically, with the increase of the feedback delay δ, the stability region of system (5) first quickly reduces to a minimum and then slowly increases.
ree representative values δ � 0.5, δ � 4, and δ � 10 are chosen, and we can see that system (5) converges the slowest when δ � 4 which is the nearest to the minimum point (see Figure 12).

Conclusion
In this paper, a fractional-order mosaic virus infection model for Jatropha curcas with farming awareness and an execution delay was studied based on the model in [2]. Compared with the studies, we not only generalized the integral order system to a fractional-order system which may describe the mechanism of the disease transmission more accurately but also investigated the Hopf bifurcation control through a time-delayed feedback controller. We mainly studied the existence of Hopf bifurcation and explored the impact of time delay, the fractional order, the feedback gain, and the feedback delay on the occurrence of bifurcation phenomenon.
By choosing the execution delay as a bifurcation parameter, Hopf bifurcation of the uncontrolled system was firstly studied. e stability of the coexistence equilibrium and the bifurcation criteria were obtained. en, by introducing a time-delayed feedback controller, the bifurcation control problem of the uncontrolled system was investigated in detail. A series of numerical simulations were performed, which not only verified our theoretical results but also revealed some specific features. Numerically, we exploited the impact of the time delay, the fractional order, the feedback gain, and the feedback delay on the occurrence of Hopf bifurcation. We found that, for both uncontrolled and controlled systems, when the execution delay is small enough (less than the critical values), the systems are stable, while they lose their stability and Hopf bifurcations occur when the execution delay exceeds the critical values. Besides, our study showed that the bifurcation values are very sensitive to the change of the fractional order. With the increase of the order, the bifurcation values of the two systems decrease rapidly, which implies that the stability region of systems becomes smaller. Furthermore, we investigated the impact of the feedback gain and the extended feedback delay on the bifurcation value for the controlled system, and we found the existence of the extreme points for the feedback gain and the extended feedback delay which can minimize the bifurcation value.
According to our study, the execution time delay can lead to the fluctuation of the population quantity of Jatropha plants, which means that the population quantity is at some unreasonable level. By introducing a timedelayed feedback controller, we can stabilize the population by harvesting or transplanting new plants according to the current and past population level (at time t and t − δ). If the population quantity of the healthy plant at t − δ is more than the current level, we harvest some plants. Otherwise, we transplant some new plants. Our numerical simulations show that there exist extreme points for the feedback gain k and the feedback delay δ which can both minimize the bifurcation value, so we can select values of the parameters k and δ which are far away from those extreme points, then the population size tends to be stabilized. Specifically, we prefer to choose the feedback gain k which is greater than the minimum point and choose the feedback delay δ which is less than the minimum point because these values are more conducive to a stable population level.
For the control of system (3), we can also add a controller to the other equations or add more than one controller at a time, and comparisons between different controllers would be very meaningful. Besides, introducing artificial impulsive control behaviors into system (3) is also interesting. We will continue these studies in our future work.