Table of Contents Author Guidelines Submit a Manuscript
Computational and Mathematical Methods in Medicine
Volume 2014, Article ID 172923, 12 pages
http://dx.doi.org/10.1155/2014/172923
Research Article

A Mathematical Model of Cancer Treatment by Radiotherapy

1School of Science, Chongqing Jiaotong University, Chongqing 400074, China
2Department of Mathematics, Hangzhou Normal University, Hangzhou, Zhejiang 310036, China
3School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 610054, China

Received 22 July 2014; Revised 14 October 2014; Accepted 25 October 2014; Published 13 November 2014

Academic Editor: Huaguang Zhang

Copyright © 2014 Zijian Liu and Chenxue Yang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

A periodic mathematical model of cancer treatment by radiotherapy is presented and studied in this paper. Conditions on the coexistence of the healthy and cancer cells are obtained. Furthermore, sufficient conditions on the existence and globally asymptotic stability of the positive periodic solution, the cancer eradication periodic solution, and the cancer win periodic solution are established. Some numerical examples are shown to verify the validity of the results. A discussion is presented for further study.

1. Introduction

Cancer is a well-known killer of humans worldwide, and its treatments are varied and sporadically successful. There are four main types of cancer treatments, which are surgery, chemotherapy, radiotherapy, and immunotherapy. In this paper, we only consider cancer treatment by radiotherapy.

Radiotherapy, as a primary treatment strategy, has been proven to be an effective tool in combating with cancer [1, 2]. Radiation therapy is a treatment procedure that uses radiation to kill malignant cells. This treatment targets rapidly reproducing cells such as those in cancer [3]. Therefore, when cancer cells are irradiated, there is a lesser effect on more slowly reproducing surrounding healthy cells. As such, the intent of this paper is to model the dynamics and interactions of healthy and cancer cells under radiation therapy.

It is an important and effective way to deeply understand the real-world problems by establishing mathematical models and analyzing their dynamical behaviors (see [49] and reference cited therein). Recently, some mathematical models that focus on cancer treatment by radiotherapy have been presented and studied ([1013]). Liu et al. in [10] focused on dynamical behaviors of normal cells that affected periodic radiation and established some conditions on the permanence and extinction of the normal and radiated cells. Moreover, they obtained criteria on the existence and global asymptotic stability of unique positive periodic solutions of the system. Belostotski in [11] presented a mathematical model to represent the interactions between healthy and cancer cells subject to radiation, where the interactions between healthy and cancer cells were viewed as competition for bodily resources. He featured four different control mechanisms of radiation delivery. They included continuous constant radiation, continuous radiation that is proportional to the instantaneous cancer concentration, continuous radiation that is proportional to the ratio of cancer to healthy cell concentration, and periodic administration of radiation. He supposed that the effect of radiation on healthy cells ideally is zero and obtained some sufficient conditions on each case that guarantee the cancer to be cure or treatment. In paper [12], Belostotski and Freedman developed and analyzed a mathematical model of cancer treatment by radiotherapy using control theory, where the radioactivity only affected the cancer cells. Later, considering the fact that the radiation also may affect the healthy cells to some extent during the radiotherapy, Freedman and Belostotski in [13] extended the previous study by perturbing the previous models. They considered four types of treatment delivery: constant, linear, feedback, and perturbed periodic deliveries. For each case, they established some sufficient conditions on the cure state and treatment state. However, paper [13] only considered the perturbed periodic radiation, although paper [12] investigated the periodic radiation, it supposed that the effect of radiation on healthy cells is zero. Hence, the study of periodic radiation under conditions that both cancer and healthy cells are affected by radiation is of major importance.

This paper is organized as follows. In Section 2, we present our model and give a basic theorem. Conditions for the coexistence of the healthy cells and cancer cells of the system are obtained in Section 3. In Section 4, we establish some sufficient conditions on the existence and globally asymptotic stability of positive periodic solution, cancer eradication periodic solution, and cancer win periodic solution. Numerical simulations are shown to verify the validity of the theorems in Section 5. Finally, we discuss our results and present some interesting problems.

2. The Model

To simplify the model, we assume that the concentrations of cancer and healthy cells exist in the same region of the organism; the administration of radiation removes a large amount of cancer cells and a small amount of healthy cells from the system. Here, the terms “large” and “small” are used as a relation to the appropriate cell population at a particular location in the organism. Radiotherapy is in fact a control mechanism on the rates of change of the concentrations of cancer and healthy cells by harvesting them.

In a given tissue, let be the concentration of healthy cells, and let be the concentration of cancer cells; then our model takes the form where and is the strategy of the radiotherapy. We assume that when (treatment stage) and when (no treatment stage) for all , where is the periodic of treatment and is the radiation treatment time. Further, in the absence of treatment ( for all ), the interactions between cancer and healthy cells were viewed as competition for bodily resources and the model was taken Lotka-Volterra competition type [1416]. During the process of cancer radiation treatment, the healthy cells are also affected. The proportion of the radiation is , ( is the ideal, but impossible to achieve in a practical scenario). are the respective proliferation coefficients, are the respective carrying capacities, and are the respective competition coefficients.

In the absence of radiation, cancer (i.e., ) wins resulting in the following conditions [14]: This yields one globally stable equilibrium at for positive initial values [14]. The following discussion of this paper will also base on the condition (2).

According to biological interpretation, we only consider the nonnegative solutions. Hence, we suppose that and ; then the following assertion is of major importance.

Theorem 1. (i) Nonnegative quadrant of is invariant for system (1). (ii) System (1) is ultimately bounded.

Proof. Let and be solutions of system (1). Here, we only analyze the healthy cells, the cancer cells can be analyzed similarly.
(i) If at some , then for all . If for all , from the first equation of system (1), for all . Hence, for any nonnegative initial values.
(ii) From the first equation of system (1) for all . According to the comparison theorem, there is a such that for all . is ultimately bounded. This completes the proof.

3. Coexistence

In this section, we will investigate the coexistence of the healthy and cancer cells. We will find that when the radiation dosage is chosen from a given interval, the cancer cells will neither grow unrestricted nor tend to zero.

To understand the model more clearly, we rewrite system (1) as follows:

Before giving the main result of this section, we firstly consider the following two-specie Lotka-Volterra competitive system: where . We have the following useful Lemma.

Lemma 2. System (6) has a unique positive equilibrium point if inequality or is satisfied. Moreover, if inequality (7) is satisfied, then the equilibrium point of system (6) is globally asymptotically stable and if inequality (8) is satisfied, is unstable.

System (6) is a well-known Lotka-Volterra competitive system. The existence and stability of the positive equilibrium point have been studied in many articles and books, for example, [17]. Here we omit the proof of it.

On the coexistence of the healthy cells and cancer cells, we have the following theorem.

Theorem 3. Assume that the following conditions hold. Then the healthy cells and cancer cells are coexistent.

Proof. From system (5) we have for all . Using the comparison theorem (see [18, 19]) we have and for all , where is the solution of the following auxiliary system: with the initial values and . By condition (9) and Lemma 2, we know that system (11) has a unique positive equilibrium point which is globally asymptotically stable. Hence, there is a such that and for all , which imply that system (1) is permanent; that is, the healthy cells and cancer cells are coexistent. This completes the proof.

Remark 4. Condition (9) implies the range of radiation dosage . In fact, from condition (9), on the one hand we have and ; on the other hand we have Combine the above two inequalities; there is However, we cannot judge the sign of the coefficient of in inequality (13) only by the conditions (2) and (9). This leads to three cases.
(a) Consider . If , then If , is inexistent.
(b) Consider . If , then . If , is inexistent.
(c) Consider . If , then . If , then . Furthermore, if , then If , is inexistent.

Remark 5. Condition (9) guarantees that the comparison system (11) has a globally asymptotically stable equilibrium point. Therefore, we can obtain that system (1) has a positive lower bound. However, if the condition is satisfied, that is, the positive equilibrium point of the comparison system (11) is a saddle point. How to choose the radiation dosage and what dynamical behavior of system (1) are still open problems.

Remark 6. Under conditions of Theorem 3, system (1) has a positive -periodic solution. The biological meanings can be understood as follows. If there is no treatment, from condition (2) we know that system (1) has a globally stable equilibrium point for any positive initial values. However, if we treat it all the time, by condition (9) and Lemma 2, we know that system (11) has a unique globally asymptotically stable positive equilibrium point . Hence, if the treatment is periodic, the solution of system (1) will tend to the positive equilibrium point during the treatment stage, and it will tend to the cancer win state during the no treatment stage. This will lead to the appearance of a periodic solution.

4. Periodic Solutions

The existence of a positive -periodic solution is guaranteed by Remark 6. In the following, we will firstly give some criteria on the existence of cancer eradication periodic solution and cancer win periodic solution of system (1). Whereafter, we will establish some conditions under which each periodic solution is globally asymptotically stable.

Firstly, let us investigate the existence of cancer eradication periodic solution of the system. Consider the following subsystem of system (1) under the case that : When , we have from the continuity of , Then when , we have hence, If system (17) has a positive -periodic solution, then we need , which leads to From expression (22) we can calculate that is equivalent to . Then we get the following result.

Theorem 7. System (1) has a cancer eradication periodic solution if the inequality is satisfied.

The existence of the cancer win periodic solution of system (1) can be analyzed similarly. Here we only show the result.

Theorem 8. System (1) has a cancer win periodic solution if the inequality is satisfied.

Up to now, we have completed the studies on the existences of the positive periodic solution, the cancer eradication periodic solution, and the cancer win periodic solution of system (1). In the following of this section, we will mainly study the global stability on each periodic solution.

Let be a positive periodic solution of system (1). Choose where is any solution of system (1). Calculating the derivative of along system (5), when , we have When , we get Consequently, if and , from (24) and (25), we have for all . By Lyapunov stability theory (see [18, 19]), the following theorem is obtained immediately.

Theorem 9. The positive periodic solution of system (1) is unique and globally asymptotically stable if the conditions and are satisfied.

On the uniqueness and global stabilities of the cancer eradication periodic solution and cancer win periodic solution, we have the following results.

Theorem 10. Assume that condition holds. Further, if are satisfied, where and , then system (1) has a unique globally asymptotically stable cancer eradication periodic solution.

Proof. The existence of the cancer eradication periodic solution has been established by Theorem 7; then we will mainly prove its uniqueness and global stability.
Let be any solution of system (1). Choose When , by condition (26) and calculating the derivative of along system (5), we have Hence, for all . According to the continuity of the solution, especially, we have .
However, when , by condition (26) and calculating the derivative of along system (5), we have Consequently, for all , where and is bounded. From the above analysis, we know that for all .
By a simple calculation and from the last inequality of condition (26), it can be obtained that as . Together with (30), we finally have as . Hence, and as . The cancer eradication periodic solution is unique and global stable. This completes the proof.

Theorem 11. Assume that condition holds. Further, if are satisfied, where . Then system (1) has a unique globally asymptotically stable cancer win periodic solution.

Proof. Let be any solution of system (1). Choose Proceed the similar analysis as Theorem 10; condition (31) can be obtained easily. Here we omit the proof of it.

Remark 12. In the discussion of the existence and global stability of the positive periodic solution, we need the conditions that and . However, from the fact we realize that is allowed in the existence and global stability of the cancer eradication periodic solution by Theorems 7 and 10 and is also allowed in the existence and global stability of the cancer win periodic solution by Theorems 8 and 11. These show that a high radiation dosage may kill the cancer cells all (Theorem 10) and it may also kill the healthy cells all (Theorem 11).

5. Numerical Illustrations

In this section, we give three groups of numerical examples to verify the validity of the three cases of periodic solutions, respectively. We consider system (1) with the following coefficients in Theorem 9: , , , , and . Obviously, condition (2) is satisfied. If the rate to the healthy cells from the radiation is chosen as , then we have and . The range of is determined by case (a) in Remark 4. Calculating it simply, we get that . We choose here. It is easy to verity that all the conditions of Theorems 3 and 9 are satisfied. Hence, when we consider hours as a treatment period, system (1) has a unique globally asymptotically stable positive -periodic solution. See Figures 13. The treatment time is chosen as hours (Figure 1), hours (Figure 2), and hours (Figure 3), respectively.

fig1
Figure 1: The dynamics of system (1) with . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable positive -periodic solution.
fig2
Figure 2: The dynamics of system (1) with . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable positive -periodic solution.
fig3
Figure 3: The dynamics of system (1) with . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable positive -periodic solution.

From Figures 13, it is interesting to observe that as the increasing of the treatment time in a periodic treatment, the concentration of the healthy cells will increase and the concentration of the cancer cells will decrease. But we cannot eradicate the cancer cells no matter how to choose the treatment time in this situation.

To show the existence and global stability of the cancer eradication periodic solution, we illustrate Theorem 10 with the coefficients in system (1) as , , . Then, and , condition (2) is satisfied. We choose and . Obviously, if we choose the periodic of treatment hours, by the fact , we have . The cancer eradication periodic solution exists by Theorem 7. See Figure 4(a) ( hours) and Figure 4(b) ( hours). From Figure 4(a), we can obtain that , and, Figure 4(b), . It is easy to verify that condition (26) is satisfied. According to Theorem 10, system (1) has a unique globally asymptotically stable cancer eradication -periodic solution. See Figures 5(a) and 5(b) ( hours) and Figures 6(a) and 6(b) ( hours).

fig4
Figure 4: (a) The time series for the cancer eradication periodic solution with and . (b) The time series for the cancer eradication periodic solution with and .
fig5
Figure 5: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer eradication -periodic solution.
fig6
Figure 6: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer eradication -periodic solution.

It can be seen from Figures 5(a) and 5(b) and Figures 6(a) and 6(b) that if all the other conditions are not changed, the increase of the treatment time in a periodic treatment will quicken the extinction of the cancer cells and decrease the concentration of the healthy cells at the same time.

As we all know, larger dosage radiation can kill cancer cells more effectively, but it also may increase the rate to the healthy cells from the radiation. We now investigate effects of the variance of the parameter to the cancer eradication periodic solution with hours and hours. Here, we also choose . For the sake of getting a cancer eradication periodic solution, we need the condition . Then we have . However, the true Theorem 10 needs condition (26), which includes that . From the fact that , then we have , which implies . Hence, here we choose and separately to investigate the influence of the variance of the to the cancer eradication periodic solution. The dynamical behavior of the system with can be seen in Figures 5(a) and 5(b).

If we take , it is easy to verify that and system (1) has a cancer eradication periodic solution. See Figure 7(a). From Figure 7(a) we obtain ; then condition (26) is satisfied. According to Theorem 10, system (1) has a unique globally asymptotically stable cancer eradication -periodic solution. See Figures 8(a) and 8(b). If is taken as , condition is also satisfied, then system (1) has a cancer eradication periodic solution. See Figure 7(b). From Figure 7(b) we obtain , then it is easy to verify that condition (26) is satisfied. According to Theorem 10, system (1) has a unique globally asymptotically stable cancer eradication -periodic solution. See Figures 9(a) and 9(b). It can be seen from Figures 5(a) and 5(b) (), Figures 8(a) and 8(b) (), and Figures 9(a) and 9(b) () that under the circumstance that the system has a cancer eradication periodic solution and all the other conditions are not changed, the increase of the radiation dosage will quicken the extinction of the cancer cells but will also decrease the concentration of the healthy cells at the same time. One can refer to the three figures; in Figure 5 with , , in Figure 8 with , , and in Figure 9 with , .

fig7
Figure 7: (a) The time series for the cancer eradication periodic solution with and . (b) The time series for the cancer eradication periodic solution with and .
fig8
Figure 8: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer eradication -periodic solution.
fig9
Figure 9: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer eradication -periodic solution.

It is also a helpful suggestion for doctors that under what situation the cancer will win the competition. Theorem 11 gives us a useful result. Let us illustrate the theorem more clearly applying the following numerical examples. Choose the coefficients , , , , , and . It is easy to verify that condition (2) is satisfied. We choose the periodic of treatment hours and the rate to the healthy cells from the radiation is . Firstly, we let the radiation dosage . From the fact that we know that , then system (1) has a cancer win periodic solution by Theorem 8. See Figure 10(a), where the treatment time hours. From Figure 10(a), we can obtain that . It is easy to verify that condition (31) is satisfied. According to Theorem 11, system (1) has a unique globally asymptotically stable cancer win -periodic solution. See Figures 11(a) and 11(b). However, by Remark 12, we know that the radiation dosage is also permitted. Hence, we choose in Figure 12. Then we have , system (1) has a cancer win periodic solution by Theorem 8. See Figure 10(b), where the treatment time hour. From Figure 10(b), we can obtain that . It is easy to verify that condition (31) is satisfied. According to Theorem 11, system (1) has a unique globally asymptotically stable cancer win -periodic solution. See Figures 12(a) and 2(b). It can be seen from Figures 11(a) and 11(b) and Figures 12(a) and 12(b) that the cancer treatment is affected by both the radiation dosage and the treatment time, if we increase the radiation dosage and decrease the treatment time at the same time, the cancer will also win the competition.

fig10
Figure 10: (a) The time series for the cancer win periodic solution with and . (b) The time series for the cancer win periodic solution with and .
fig11
Figure 11: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer win -periodic solution.
fig12
Figure 12: The dynamics of system (1) with and . (a) The time series for healthy cells and cancer cells . (b) The phase of healthy cells and cancer cells . Obviously, system (1) has a globally stable cancer win -periodic solution.

Throughout Figures 112, we choose the initial values and . Also note that the choice of the values in the theorems are based on the range determined by [3], but they do not come from any real cell populations.

6. Conclusion and Discussion

In this paper, we employed a pair of ordinary differential equations to model the dynamics between the healthy cells and cancer cells for the cancer treatment by radiotherapy. We separated the treatment into two stages: treatment stage and recovery stage (no treatment stage). During the treatment stage, the radiation harvesting amount is to the cancer cells and to the healthy cells. However, during the recovery stage, the model has taken the most basic Lotka-Volterra competition type. We gave the range of radiation dosage under the following three results: the healthy cells and cancer cells are coexist, the cancer eradication periodic solution is globally stable, and the cancer win periodic solution is globally stable. Note that, in the discussion on the global stability of the cancer eradication periodic solution, we showed the relationship between the radiation dosage and the treatment time ; that is, and , which are also given during the discussion on the cancer win periodic solution (). However, during the discussion on the coexistence of the healthy cells and cancer cells, we only give the range of the radiation dosage , which shows that it fits all the treatment time . This is a flaw, which is caused by the analysis methods in this paper. We will improve our analysis techniques and make a further study in the future.

The cancer treatment model discussed in this paper is only based on one treatment measure. It may be more effective for the cancer treatment if we add a medication during the recovery stage, which is still an open problem and we will carry out the research in the further work.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This research was supported by the National Natural Science Foundation of China (11401060) and Zhejiang Provincial Natural Science Foundation of China (LQ13A010023).

References

  1. C. A. Perez and J. A. Purdy, Levitt and Tapley's Technological Basis of Radiation Tehrapy: Clinical Applications, Williams and Wilkins, Philadelphia, Pa, USA, 3rd edition, 1999.
  2. R. L. Souhami, I. Tannock, P. Hohenberger, and J.-C. Horiot, Oxford Textbook of Oncology, vol. 1, Oxford University Press, Oxford, UK, 2nd edition, 2002.
  3. G. G. Steel, The Biological Basis of Radiotherapy, Elsevier, 1989.
  4. Z. Liu, M. Fan, and L. Chen, “Globally asymptotic stability in two periodic delayed competitive systems,” Applied Mathematics and Computation, vol. 197, no. 1, pp. 271–287, 2008. View at Publisher · View at Google Scholar · View at MathSciNet · View at Scopus
  5. H. Zhang, Z. Wang, and D. Liu, “Global asymptotic stability of recurrent neural networks with multiple time-varying delays,” IEEE Transactions on Neural Networks, vol. 19, no. 5, pp. 855–873, 2008. View at Publisher · View at Google Scholar · View at Scopus
  6. L. V. Hien, T. T. Loan, B. T. Trang, and H. Trinh, “Existence and global asymptotic stability of positive periodic solution of delayed Cohen-Grossberg neural networks,” Applied Mathematics and Computation, vol. 240, no. 1, pp. 200–212, 2014. View at Publisher · View at Google Scholar · View at MathSciNet · View at Scopus
  7. H. Zhang, W. Huang, Z. Wang, and T. Chai, “Adaptive synchronization between two different chaotic systems with unknown parameters,” Physics Letters A, vol. 350, no. 5-6, pp. 363–366, 2006. View at Publisher · View at Google Scholar · View at Scopus
  8. H. Gu, H. Jiang, and Z. Teng, “Existence and globally exponential stability of periodic solution of BAM neural networks with impulses and recent-history distributed delays,” Neurocomputing, vol. 71, no. 4–6, pp. 813–822, 2008. View at Publisher · View at Google Scholar · View at Scopus
  9. H. Zhang, Z. Wang, and D. Liu, “Robust exponential stability of recurrent neural networks with multiple time-varying delays,” IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 54, no. 8, pp. 730–734, 2007. View at Publisher · View at Google Scholar · View at Scopus
  10. Z. Liu, S. Zhong, C. Yin, and W. Chen, “Permanence, extinction and periodic solutions in a mathematical model of cell populations affected by periodic radiation,” Applied Mathematics Letters, vol. 24, no. 10, pp. 1745–1750, 2011. View at Publisher · View at Google Scholar · View at MathSciNet · View at Scopus
  11. G. Belostotski, A control theory model for cancer treatment by radiotherapy [M.S. thesis], University of Alberta, 2004.
  12. G. Belostotski and H. I. Freedman, “A control theory model for cancer treatment by radiotherapy 1: no healthy cell damage,” International Journal of Applied Mechanics, vol. 25, no. 4, pp. 447–480, 2005. View at Google Scholar
  13. H. I. Freedman and G. Belostotski, “Perturbed models for cancer treatment by radiotherapy,” Differential Equations and Dynamical Systems, vol. 17, no. 1-2, pp. 115–133, 2009. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet · View at Scopus
  14. H. I. Freedman, Deterministic Mathematical Models in Population Ecology, Marcel Dekker, New York, NY, USA, 1980. View at MathSciNet
  15. F. Nani and H. I. Freedman, “A mathematical model of cancer treatment by immunotherapy,” Mathematical Biosciences, vol. 163, no. 2, pp. 159–199, 2000. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet · View at Scopus
  16. S. T. Pinho, H. I. Freedman, and F. Nani, “A chemotherapy model for the treatment of cancer with metastasis,” Mathematical and Computer Modelling, vol. 36, no. 7-8, pp. 773–803, 2002. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet · View at Scopus
  17. Z. Ma, Mathematical Modeling and Research on the Population Ecology, Anhui Educational Press, Hefei, Hefei, 1996, (Chinese).
  18. V. Lakshmikantham, V. M. Matrosov, and S. Sivasundaram, Vector Lyapunov Functions and Stability Analysis of Nonlinear Systems, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1991. View at MathSciNet
  19. T. Yoshizawa, Stability Theory and the Existence of Periodic Solutions and Almost Periodic Solutions, Springer, New York, NY, USA, 1975.