Existence and uniqueness of a positive steady state solution for a logistic system of dierential dierence equations

V ctor Padrn o . Departamento de Matemticas a Facultad de Ciencias Universidad de Los Andes Mrida 5101, Venezuela e

Abstract We consider a logistic system of dierential dierence equations that arises as a population model for the dispersal of a single species in a patchy landscape. Using monotone methods we prove the existence and uniqueness of a positive steady state solution. Since the trivial solution is unstable, the theory of cooperative and irreducible systems guarantee that the generic solution converges to a single equilibrium point. Key words : monotone matrices, irreducible matrix, maximum principle, patchy landscape.
Current address: School of Mathematics, University of Minnesota, 127 Vincent Hall, 206 Church Street S.E., Minneapolis, Minnesota, 55455, padro005@umn.edu

POSITIVE STEADY STATES

Running title : POSITIVE STEADY STATES. AMS(MOS) subject classication: 92D25, 34A34

Introduction

We consider the system of dierential equations

ui (t) =
jIi

[dij uj (t) dji ui (t)] + ui (t)Fi (ui (t)), i = 1, . . . , n, t 0.

(1)

Here dij are positive constants, and Ii is a nonempty subset of {1, 2, . . . , n} such that j Ii implies i Ij , i, j = 1, . . . , n. We also assume that there is not a non-empty and proper subset I {1, 2, . . . , n} such that Ii I for all

i I. This condition on Ii accounts for assuming that the matrix A = (aij ) of the linear part of (1), dij , 0, aij =

i = j and j Ii , i = j and j Ii , / dij , i = j, jIi

(2)

is irreducible. The functions Fi (s), i = 1, . . . , n, verify the following hypothesis: Hypotheses 1 1. Fi is continuously dierentiable.

2. ri := Fi (0) > 0, and i Fi (s) i for given positive constants i , i , and any s 0.

POSITIVE STEADY STATES

This problem arises as a population model for a single species which is composed of several habitats connected by linear migration rates and having logistic growth. In this case ui (t) is the population in habitat i at time t; the coecients dij are the rates at which the individuals migrate from habitat j to habitat i; and Fi represents the net rate of population supply at habitat i. The second condition in Hypothesis 1 implies that there exists 0 < Ki < , such that fi is positive in (0, Ki ) and negative in (Ki , +), for i = 1, . . . , n. The value Ki > 0 is called the carrying capacity of the population because it represents the population size that available resources can continue to support. The value ri > 0 is called the intrinsic growth rate and represents the per capita growth rate achieved if the population size were small enough to ensure negligible resource limitations. For the standard logistic growth, s ). introduced by P.F. Verhulst [7], Fi (s) = ri (1 Ki Continuous time models with multi-patch formulation have also been proposed in the study of the spatial dynamics of epidemics (see [1] and the references therein). The rst condition in Hypothesis 1 is technical and is needed together with the second condition to ensure, given initial data, the existence and uniqueness of solutions for (1) globally dened for t 0, c.f. [3]. Since dij > 0 and the matrix A is irreducible, (1) is a cooperative and irreducible system in I n . This implies that the set C := { I n : limt u(t; ) exists } of R+ R+ convergent points of (1) contains an open and dense subset of I n , c.f. [5, R+ Theorem 4.1.2, page 57]. Here, u(t; ) denotes the solution of (1) such that

POSITIVE STEADY STATES

u(0; ) = . Thus the dynamics of (1) is largely determined by its equilibria. It is clear that u 0 is a steady state solution of (1). Since Fi (0) > 0, i = 1, . . . , n, it follows that u 0 is unstable. Hence, by the above argument there exists at least one nontrivial equilibria u such that limt u(t; ) = u for some in an open subset of C. In fact it is not dicult to see that when dij Kj =
jIi jIi

dji Ki , i = 1, . . . , n,

(3)

the only nontrivial equilibrium of (1) is u = (K1 , K2 , . . . , Kn ) (see [4, Theo rem 1]). The purpose of this paper is to prove the following theorem regarding the existence and uniqueness of a nontrivial steady state for (1). Theorem 1 There exists one and only one positive steady state solution for (1). Theorem 1 implies that the generic solution of (1) converges to a single equilibrium point. This result is well known for reaction-advection-diusion equations (see [2, Proposition 3.3, page 148]). Since (1) can be seen as a spatially discrete model analogous to an advection-diusion equation, the same result was expected to be true for (1). Nevertheless, to the best of our knowledge, its proof has not yet appeared in the literature. In order to stress the similaritt of (1) to a reaction-advection-diusion equation, we present a proof that is based on the corresponding proof for the PDE model.

POSITIVE STEADY STATES

Preliminary Results

Consider the linear system dij (xj xi ) + bi xi = fi , i = 1, . . . , n.

jIi

(4)

Here bi , and fi are given constants. Theorem 2 (Strong Maximum Principle) Suppose that bi 0, i = 1, . . . , n. Let x I n be a solution of (4), m := min xi , and M := max xi . If R
1in 1in

fi 0, i = 1, . . . , n, and m 0 (resp. fi 0, i = 1, . . . , n, and M 0), then xi = m (resp. xi = M ), and fi = mbi = 0, i = 1, . . . , n. In particular, if fi 0 (resp. fi 0), i = 1, . . . , n, and f 0 or b 0, then m 0 (resp. M 0). Remark 1 This theorem is a particular version of more general maximum principles for monotone matrices (see [6, Theorem 2]). We will give an independent proof. Proof. Suppose that fi 0, i = 1, . . . , n and m 0, (the case that fi

0, i = 1, . . . , n and M 0 can be solved similarly). Let I := {i : xi = m}. Clearly, I is non-empty. Suppose that I {1, . . . , n}. Hence, by our original I c is non-empty.

assumption on the sets Ii s, there exists i I such that Ii

Here, I c denotes the complement of I over the set {1, . . . , n}. For such i we

POSITIVE STEADY STATES have fi =

jIi

dij (xj xi ) + bi xi dij (xj m) + bi m

Ic

=
jIi

> 0. which is a contradiction. This proves that xi = m, i = 1, . . . , n. It follows, by introducing this value for xi in (4), that fi = mbi = 0, i = 1, . . . , n. We will apply the Strong Maximum Principle to obtain a monotone iteration scheme for constructing solutions to the general nonlinear system dij (xj xi ) + ei xi + fi (xi ) = 0, i = 1, . . . , n.
jIi

(5)

Here fi , i = 1, . . . , n are given functions. We will say that x0 is an upper solution of (5) if x0 satises dij (x0 x0 ) + ei xi + fi (x0 ) 0, i = 1, . . . , n. j i i
jIi

(6)

We will also assume that x0 is not a solution. Similarly we dene lower solution by interchanging the order of the inequalities in (6). Theorem 3 Suppose that fi C 1 (R) for all i, and that x0 and y 0 are re0 spectively upper and lower solutions of (5) with yi x0 , i = 1, . . . , n. Then i 0 there exist solutions x and y of (5) such that yi yi xi x0 , i = 1, . . . , n. i

Moreover, x and y are, respectively, maximal and minimal solutions of (5), in

0 the sense that if x is any solution of (5) such that yi xi x0 , i = 1, . . . , n i

then yi xi xi , i = 1, . . . , n.

POSITIVE STEADY STATES Proof.

0 Let m := min yi and M := max x0 . Let k > 0 be a constant i 0in 0in

such that k > ei , i = 1, . . . , n, and fi (s) + k > 0, 0 i n, s [m, M ]. Dene the application T : I n I n by = T where is the solution R R of the problem dij (j i ) + (ei k)i = (fi (i ) + ki ), i = 1, . . . , n.
jIi

(7)

The matrix of linear system (7) is of the form A kI where A is the matrix given in (2). By Gershgorins Theorem we know that A has nonpositive eigenvalues. Hence, since ei k < 0, i = 1, . . . , n, the matrix A kI is invertible, and T is well dened. We will see that T is a monotonic application in the sense that if 1 2 then T 1 T 2 , provided that m 1 , 2 M . Here 1 2 means
1 2 1 2 i i , i = 1, . . . , n, and 1 < 2 means i < i , i = 1, . . . , n. To see this

let i = T i , i = 1, 2. Then

2 1 dij ((j j ) (i2 i1 )) + (ei k)(i2 i1 ) jIi

(8) i = 1, . . . , n.

2 (fi (i )

1 fi (i )

2 k(i

1 i )),

If we assume that 1 2 , then by the choice of k we have

2 1 dij ((j j ) (i2 i1 )) + (ei k)(i2 i1 ) 0, i = 1, . . . , n. jIi

(9)

This implies that 1 < 2 . It is clear that 1 2 . Suppose that there exists i such that i1 = i2 . By the Strong Maximum Principle 1 2 . In

POSITIVE STEADY STATES

this case, the left hand side of (8) is zero. It follows by the choice of k that 1 2 . Hence, T is monotonic in a stronger sense, i.e. if 1 < 2 then T 1 < T 2 . Let us see now that T < if is an upper solution of (5). Let = T , then

dij ((j j ) (i i )) + (ei k)(i i )

jIi

= (
jIi

dij (j i ) + fi (i )) 0, i = 1, . . . , n.

(10)

Again, by the Strong Maximum Principle < , otherwise . This last option is not possible since by the denition of T , would be a solution of (5) and we assumed that this was not the case. This observations allow us to dened inductively two sequences {xn } and {y n } by letting (x1 , y 1 ) := (T x0 , T y 0 ) and (xn , y n ) := (T xn1 , T y n1 ) for n > 1. Since x0 is an upper solution, x1 = T x0 < x0 , and by the monotonicity of T , T x1 < T x0 = x1 . Hence xn1 > xn for each n. Similarly, y n > y n1 for each n. Also, since x0 y 0 , we obtain by induction that xn y n . This allow us to conclude that {xn } is a decreasing sequence bounded below by y 0 . Hence xi = lim xn , i = 1, . . . , n, i
n

exists.

POSITIVE STEADY STATES Now, x = lim xn = lim T xn1 = T lim xn1 = T x.

n n n

Then x is solution of (5). Similarly we can construct a solution y of (5) such that y 0 y x x0 . Suppose now that x is another solution of (5) such that y 0 x x0 . This implies that x = T x < T x0 = x1 . By induction x < xn for all n. Hence x x. Similarly, one shows that y x. This nishes the proof of the theorem.

Proof of the main Theorem

The proof of Theorem 1 relies in the following lemmas Lemma 1 If x is a solution of (5) with fi (s) := sFi (s), then xi a , i =
r +r 1, . . . , n, where a := max1in { ii , eii i }.

Proof.
1in

Suppose that there exists i such that xi > a . Let i0 such that

xi0 = max xi > a . Hence ei0 xi0 + fi0 (xi0 ) < 0. It follows that di0 j (xj xi0 ) + ei0 xi0 + fi0 (xi0 ) ei0 xi0 + fi0 (xi0 )
jIi0

< 0. This is a contradiction which proves the lemma. The steady state solutions of (1) are solutions of a system of the form dij (xj xi ) + ei xi + xi Fi (xi ) = 0, i = 1, . . . , n.
jIi

(11)

POSITIVE STEADY STATES Lemma 2 If the principal eigenvalue 1 is positive in the problem dij (j i ) + ei i + ri i = i , i = 1, . . . , n,
jIi

10

(12)

where ri = Fi (0), then system (11) has one and only one positive solution. By a positive solution of (11) we mean a solution x I n such that R xi > 0, i = 1, . . . , n. Proof. To prove the existence of a positive solution of (11) we rst notice

r +r that a := max1in { ii , eii i } is an upper solution of (11).

Let a positive eigenvector corresponding to the main eigenvalue 1 of (12). It follows that satises the equation dij ( j i ) + ei i + i Fi ( i ) = i (1 ri + Fi ( i )), i = 1, . . . , n.
jIi

Since 1 + Fi ( i ) Fi (0) 1 i i we obtain, by choosing

> 0 small

enough, that is a positive lower solution of (11). We can assume without loss of generality that i < a , i = 1, . . . , n. It follows by Theorem 3 that (11) has a positive solution. Let x be the maximal solution of (11) given by Theorem 3 for the upper solution a . Suppose that x is another positive solution of (11). By Lemma 1 xi a , i = 1, . . . , n. Since x is a maximal solution, it follows that 0 xi xi , i = 1, . . . , n. We will show that this implies that x = x. Since x is a positive solution of (11) it follows that the principal eigenvalue 1 is zero for the problem

POSITIVE STEADY STATES

11

dij (j i ) + ei i + Fi (xi )i = i , i = 1, . . . , n,
jIi

(13)

Similarly, since x is a positive solution of (11) it follows that the principal eigenvalue 1 is zero for the problem

dij (j i ) + ei i + Fi (i )i = i , i = 1, . . . , n. x
jIi

(14)

Assume that there exists i such that xi < xi . This implies that Fi (i ) < x Fi (xi ). Hence, applying the monotonicity property of the principal eigenvalue of nonnegative irreducible matrices to problems (13)-(14), it follows that 1 < 1 . This is a contradiction, since 1 = 1 = 0. We conclude that x = x. This nished the proof of the theorem. Proof of Theorem 1. Since the principal eigenvalue of the matrix A is zero then the principal eigenvalue of the problem [dij j dji i ] + ri i = i , i = 1, . . . , n,
jIi

(15)

is positive. The theorem follows by applying Lemma 2 to (1) with ei =

jIi (dij

dji ).

References
[1] J. Arino, J.R. Davis, D. Harley, R. Jordan, J.M. Miller, and P. Van Den Driessche, A multi-species model with spatial dynamics, Mathematical Medicine and Biology 22 (2005), 129142.

POSITIVE STEADY STATES

12

[2] R.S. Cantrell, C. Cosner, Spatial ecology via reaction-diusion equations, Wiley, 2003. [3] V. Padrn, M.C. Trevisan, Eect of Aggregating Behavior on Population o Recovery on a set of habitat islands, Math. Biosc. 165 (2000), no. 1, 63-78. [4] V. Padrn, M.C. Trevisan, Environmentally Induced Dispersal Under o Heterogeneous Logistic Growth, Math. Biosc 199 (2006) 160174. [5] H. Smith, Monotone dynamical systems: an introduction to the theory of competitive and cooperative systems, American Mathematical Society, SURV/41, 1995. [6] G. Stoyan, Maximum principles for monotone matrices, Lin. Alg. Appl. 78 (1986) 147161. [7] P.F. Verhulst, Notice sur la loi que la population suit dans son accroissement, Corr. Math. et Phys. 10 (1838) 113121.