1. Introduction
Advanced-type differential equations are used to describe phenomena in which the evolution of the system depends on both present and future time. The possibility of introducing an advanced argument into an equation to take into account future influences that may actually affect the present makes such equations a useful tool in various economic problems, in the mathematics of networks, and in optimization, as well as in modeling engineering problems such as in electrical power systems, materials, and energy (see [
1]).
Oscillation theory represents one of the basic research fields in both physics and mathematics that deals with the behavior of systems that exhibit periodic motions such as in population dynamics, rhythmic beating of the human heart, and electrical circuits. Also, the study of oscillations is important not only in such natural phenomena as the movement of tides and planetary motion, but also in technological systems such as mechanical engineering, control systems, and signal processing (see [
2,
3]). It is also a well-established area of research on its own.
Delay differential equations of various orders, that is, equations containing both present and past history, have long been the focus of scientists and researchers, and numerous oscillation results have emerged as a consequence; see for example, the monograph [
2] and the references therein. However, the ability of advanced-type differential equations to describe and analyze complex phenomena in the real world makes their importance increasingly significant. They are important in medicine, where advanced-type differential equations help to understand drug interactions within the body, and in the study of disease dynamics. They are also useful in weather forecasting, studying ocean movement, and predicting earthquakes.
In view of the variety of applications described above, it is easy to see why the qualitative theory of such equations has gained increased momentum in recent years, including the study of the oscillatory and asymptotic behavior of solutions of third-order functional differential (see the monograph [
3] and the references cited therein). In the references [
4,
5,
6,
7,
8,
9], the authors studied the oscillatory behavior of third order delay differential equations, while in [
10,
11,
12,
13] advanced-type equations are considered. Moreover in [
14,
15,
16,
17], the authors examined similar problems for differential equations with deviating arguments, and see [
18,
19,
20,
21] for differential equations with mixed arguments.
In this paper, we wish to develop conditions for the oscillation of the advanced Thomas–Fermi-type functional differential equation
      
      where
      
Without further mention, we assume:
Hypothesis 1. , , and ;
 Hypothesis 2. , , and  for ;
 Hypothesis 3. α is the ratio of odd positive integers.
 By a 
solution y of (
1), we mean a function 
, 
, where 
 and 
 are continuously differentiable and satisfy (
1). We restrict our attention to solutions 
 of (
1) that are continuable; that is, they satisfy 
 for any 
. A solution is 
oscillatory if it possesses a sequence of zeros that tends to infinity, and is 
nonoscillatory otherwise.
From a review of the literature, we see that most papers on qualitative theory of functional differential equations are devoted to Emden–Fowler -type equations, viz.,
      
      under one or more of the conditions
      
For example, in [
5], the authors considered the Emden–Fowler-type delay differential equation
      
      In [
4], the authors studied this equation with 
, and in [
6,
8,
9] with 
. As for Thomas–Fermi-type equations, in [
15,
16,
17], the equation
      
      is analyzed, and in [
5],
      
      where 
 or 
 is studied. In [
12], the authors considered the advanced-type differential equation
      
      in [
11] the equation
      
      and in [
13]
      
      all with 
.
Mixed-type equations such as
      
      (see [
7]) and
      
      (see [
20]) under the conditions
      
      have also been considered.
In [
22] (also see [
9]), the author introduced the terminology of “canonical, noncanonical, and semicanonical” equations (depending on the behavior of 
 as 
 for 
) for third-order Emden–Fowler-type equations. There does not seem to be any reason not to introduce such terminology for the Thomas–Fermi Equation (
1) above. Therefore, we say that (
1) is in canonical form (C) if (
2) holds, in noncanonical form (NC) if (
5) holds, and in semicanonical form (
) if (
3) holds or semicanonical form (
) if (
4) holds.
Even though Thomas–Fermi-type nonlinear differential equations appear in a number of applications (see, for example [
2,
6]), oscillation theory for such equations has received far less attention than have Emden–Fowler-type equations. The first detailed investigation in this direction appears that in the monograph of Chanturia and Kiguradze [
15]. Their results are especially concerned with what is known as Property B for solutions of the linear equation
      
      By Property B of (
7), we mean that every nonoscillatory solution 
y is strongly increasing; that is, 
 as 
 for 
. Oscillation theory for equations related to (
1) has been investigated in [
4,
12,
18,
20]. For example, in [
18], the authors discussed the oscillation and asymptotic behavior of (
1) if (
2) holds, and in [
10,
14,
23], the authors established criteria for Property B to hold as well as finding criteria for the oscillation of the mixed type Equation (
6) with 
 and 
, and condition (
2) or (
4) holding. Since this equation is of semi-canonical type, it has three classes of nonoscillatory solutions, and to obtain oscillation criteria, they eliminated these three types of nonoscillatory solutions. Their methods needed the presence of both delay and advanced parts.
The point here is to obtain oscillation criteria for Equation (
1); we achieve this by first transforming the semi-canonical Equation (
1) into canonical form. This approach significantly simplifies the study of Equation (
1) in the sense that the number of classes of nonoscillatory solutions are reduced from three to two without assuming any additional conditions. Then, we employ integral averaging and comparisons to obtain oscillation criteria for (
1). Some examples are given to demonstrate the scope of our main results.
  2. Main Results
For simplicity, we define the following quantities:
Theorem 1. The operator  can be written in the canonical form  Proof.  The proof of the theorem can be found from Theorem 2 in [
9]; hence, the details are omitted.    □
 Based on Theorem 1, we can rewrite Equation (
1) as the equivalent canonical equation
      
      with
      
      where 
, 
, 
, 
, and
      
      and
      
From the above discussion, we immediately obtain the following results.
Corollary 1. The semi-canonical nonlinear differential Equation (
1) 
possesses a solution  if and only if  is a solution of the canonical Equation (
8).
  Corollary 2. The semi-canonical nonlinear differential Equation (
1) 
has an eventually positive solution if and only if the canonical Equation (8) has an eventually positive solution.  Since the Equation (
8) is of the canonical type, it is possible to use Kiguradze’s lemma [
15] to obtain the following structure of its possible positive nonoscillatory solutions.
Lemma 1. Assume that z is an eventually positive solution of (
8). 
Then,oreventually.  Following [
5,
15,
23], we say that Equation (
8) has Property B if every nonoscillatory solution of (
8) belongs to the class 
.
Next, we derive some useful monotonic properties of nonoscillatory solutions that will be employed in proving our main results. For simplicity and convenience, we let
      
      and
      
      where 
 is appropriately large enough.
Lemma 2. Let  be a positive solution of (
8). 
Then,  is increasing and  Proof.  Choose 
 such that 
 is a positive solution of (
8) belonging to 
 for all 
, for some 
 large enough. From the definition of the class 
, it is easy to see that 
 and 
 are increasing for all 
. Now,
        
        which implies that 
 tends to ∞ as 
. Since 
 is also increasing, it is clear that 
 also tends to ∞ as 
. Again from (2.2), we have
        
        Using the fact that 
 as 
 in the last inequality, we see that
        
        This implies that
        
        Thus, 
 is eventually increasing, and this completes the proof.    □
 Lemma 3. Let  be a positive solution of (
8). 
Then,  is eventually decreasing.  Proof.  Let 
 be an eventually positive solution of (
8) that belongs to 
 for all 
. Then, 
 is decreasing, and so we have
        
        This implies that
        
        Hence, 
 is eventually decreasing, and this proves the lemma.    □
 Remark 1. Lemma 2 is new, and Lemma 3 is an extension of [
18] (Lemma 2.3) for the semi-canonical Equation (
1) once it has been transformed to canonical form.
 Now, we present criteria for the class  to be empty.
Theorem 2. Assume thatIfwhere  if  and  if , then the class  is empty.  Proof.  Assume, to the contrary, that 
 is a positive solution of (
8) belonging to the class 
 for all 
. Integrating (
8) from 
t to ∞ twice yields
        
        where we have used the monotonicity of 
. Integrating the last inequality from 
 to 
t and then interchanging the order of the integrations, we obtain
        
Therefore,
        
Using the fact that 
 is increasing and 
 is decreasing, we have
        
That is,
        
If 
 in (
13), we obtain a contradiction to (
12). From (
11), we see that 
 as 
, so taking the lim sup as 
 on both sides of (
13), we are again led to a contradiction to (
12) for 
. This finishes the proof.    □
 Theorem 3. Let  and assume thatIf then the class  is empty.  Proof.  Assume that 
 is a positive solution of (
8) belonging to 
 for all 
. By Lemma 3, the function 
 is decreasing. To show
        
        assume that 
. Then,
        
        for 
, and so
        
Integrating (
8) twice gives
        
This contradicts assumption (
14), and so we conclude that (
16) holds. Now if we set
        
        then (
13) implies
        
Taking the lim sup as 
 and using (
16) gives a contradiction to (
15). This proves the theorem.    □
 Remark 2. From Theorems 2 and 3, we conclude that Equation (
8) has Property B. To obtain an oscillation result by making use of theses two theorems, we would need to eliminate this possibility, that is, we would need to show that 
 is empty.
 Theorem 4. If  andwhere  if  and  if  then .  Proof.  To show that 
 is empty, assume that 
 is a positive solution of (
8) that belongs to 
 for all 
. Integrating (
8) from 
s to 
t and applying Lemma 2 gives
        
Integrating this, we obtain
        
Another integration yields
        
Setting 
 gives
        
For 
 in (
19), we obtain a contradiction to (
18). If 
, then since Lemma 2 implies 
 as 
, we again have a contradiction to (
18). The proof of the theorem is now complete.    □
 Next, we present another criterion to eliminate the class .
Theorem 5. Assume that there exist nondecreasing functions ,  such that  for all . If the first-order advanced differential equationis oscillatory, then the class  is empty.  Proof.  Assume, to the contrary, that 
 is a positive solution of (
8) belonging to 
 for all 
. For 
,
        
        so by letting 
 and 
, we obtain
        
Using this in (
8) yields
        
Let 
, 
; we see that
        
Also, similar to (
21), we have
        
        for 
, so letting 
 and 
,
        
If we let 
 and use (
23) in (
22), we have
        
Thus, 
 is a positive solution of the above inequality, and by ([
24], Corollary 2.4.2), we see that the corresponding Equation (
20) also has a positive solution. This contradiction ends the proof.    □
 For 
, using Theorem 1 (ii) in [
7], we obtain the following corollary.
Corollary 3. Let  and assume that there exists nondecreasing functions ,  such that  for all . Ifthen the class  is empty.  Corollary 4. Let  and assume that there exist nondecreasing functions ,  such that  for all . Ifthen the class  is empty.  Proof.  Proceeding as in the proof of Theorem 5, we arrive at (
24). Since 
 is increasing and 
, from (
24) we obtain
        
An integration gives
        
        which contradicts (
26) and completes the proof.    □
 Combining the criteria obtained for both of the classes 
 and 
 to be empty, we obtain the following criteria for the oscillation of Equation (
1).
Theorem 6. Let  or . If (
11), (
12), 
and (
18) 
hold, then Equation (
1) 
is oscillatory.  Proof.  If 
 is a positive solution of (
1) for all 
, then by Corollary 1, we see that 
 is a positive solution of (
8) and by Lemma 1, 
 or 
 for all 
. In view of Theorem 2, we see that the class 
 is empty and by Theorem 4, we have the class 
 is empty. Therefore, we conclude that Equation (
8) is oscillatory, which in turn implies (
1) is oscillatory. This proves the theorem.    □
 Theorem 7. Let  (). If all conditions of Theorem 2 and Corollary 3 (Corollary 4) hold, then Equation (
1) 
is oscillatory.  Proof.  The proof is similar to that of Theorem 6, and so the details are omitted.    □