Abstract

Using the comparison theorem, the two-point oscillation for linear differential equation with damping term 𝑦 ξ…ž ξ…ž + ( 𝑓 ( π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) 𝛼 ) 𝑦 ξ…ž + ( 𝑔 ( π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) 𝛽 ) 𝑦 = 0 is considered, where 𝛼 , 𝛽 > 0 ; 𝑓 ( π‘₯ ) , 𝑔 ( π‘₯ ) > 0 , and 𝑓 ( π‘₯ ) , 𝑔 ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) , 𝐼 = ( 0 , 1 ) . Results are obtained that 0 < 𝛼 < 3 / 2 , 𝛽 > 3 or 𝛼 > 3 / 2 , 𝛽 > 2 𝛼 imply the two-point oscillation of the equation.

1. Introduction

Under the solution 𝑦 ( π‘₯ ) of a differential equation appearing in the paper, we mean a function 𝑦 = 𝑦 ( π‘₯ ) such that 𝑦 ∈ 𝐢 2 ( 𝐼 ) . Here, we allow that 𝑦 βˆ‰ 𝐢 ( 𝐼 ) .

Recently, two-point oscillation of the differential equations caused the concern of many scholars ([1, 2]). In paper [1], PaΕ‘iΔ‡ and Wong construct the equation: 𝑦 ξ…ž ξ…ž +  1 2 𝑆 ξ€· π‘ž ξ…ž ξ€Έ ξ€· π‘ž ( π‘₯ ) + ξ…ž ξ€Έ 2 ξ‚„ ( π‘₯ ) 𝑦 = 0 , π‘₯ ∈ 𝐼 , ( 1 . 1 ) where 𝐼 = ( 0 , 1 ) , π‘ž ∈ 𝐢 3 ( 𝐼 ) , | π‘ž ( 0 + ) | = | π‘ž ( 1 βˆ’ ) | = + ∞ , | π‘ž ξ…ž ( 0 + ) | = | π‘ž ξ…ž ( 1 βˆ’ ) | = + ∞ , π‘ž ξ…ž ( π‘₯ ) < 0 , π‘₯ ∈ 𝐼 , 𝑆 ( π‘ž ξ…ž ) ∈ 𝐢 ( 𝐼 ) , 𝑆 ( π‘ž ξ…ž ) ( π‘₯ ) = ( π‘ž ξ…ž ξ…ž ξ…ž ( π‘₯ ) / π‘ž ξ…ž ( π‘₯ ) ) βˆ’ ( 3 / 2 ) [ π‘ž ξ…ž ξ…ž ( π‘₯ ) / π‘ž ξ…ž ( π‘₯ ) ] 2 , they study the following equation: 𝑦 ξ…ž ξ…ž + 𝑐 ( π‘₯ ) ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 𝜎 𝑦 = 0 , π‘₯ ∈ 𝐼 , ( 1 . 2 ) by comparison theorem (where 𝑐 ( π‘₯ ) > 0 , 𝑐 ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) ), and they obtain that when 𝜎 > 2 , (1.2) is two-point oscillatory.

In this paper, we construct the following equation with damping: ξ€· 𝑝 ξ…ž ( π‘₯ ) 𝑦 ξ…ž ξ€Έ ξ…ž βˆ’ 2 𝑝 ξ…ž ξ…ž ( π‘₯ ) 𝑦 ξ…ž + ξ€· 𝑝 ξ…ž ξ€Έ ( π‘₯ ) 3 𝑦 = 0 , π‘₯ ∈ 𝐼 , ( 1 . 3 ) where 𝑝 ( π‘₯ ) ∈ 𝐢 2 ( 𝐼 ) and | | | | = | | | | 𝑝 ( π‘₯ ) ( 0 + ) 𝑝 ( π‘₯ ) ( 1 βˆ’ ) = ∞ . ( 1 . 4 ) we study the two-point oscillation of the following damped equation by comparison theorem 𝑦 ξ…ž ξ…ž + 𝑓 ( π‘₯ ) ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 𝛼 𝑦 ξ…ž + 𝑔 ( π‘₯ ) ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 𝛽 𝑦 = 0 , ( 1 . 5 ) where π‘₯ ∈ 𝐼 , 𝛼 , 𝛽 > 0 ; 𝑓 ( π‘₯ ) , 𝑔 ( π‘₯ ) > 0 , 𝑓 ( π‘₯ ) , 𝑔 ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) , the result we obtained is new, and it continues the results obtained in [1].

2. Two-Point Oscillation of (1.3)

Definition 2.1. A function 𝑦 = 𝑦 ( π‘₯ ) , 𝑦 ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) is said to be two-point oscillation on the interval 𝐼 , if there exist a decreasing sequence π‘Ž ( π‘˜ ) ∈ 𝐼 and an increasing sequence 𝑏 ( π‘˜ ) ∈ 𝐼 of consecutive zeros of 𝑦 ( π‘₯ ) such that π‘Ž ( π‘˜ ) β†˜ 0 and 𝑏 ( π‘˜ ) β†— 1 .

Definition 2.2. A linear differential equation is said to be two-point oscillation on 𝐼 if all its nontrivial solutions 𝑦 = 𝑦 ( π‘₯ ) , 𝑦 ( π‘₯ ) ∈ 𝐢 2 ( 𝐼 ) are two-point oscillatory on 𝐼 .

By Sturm separation theorem, all nontrivial solutions of a linear differential equation are two-point oscillatory if there is a nontrivial solution is two-point oscillatory on 𝐼 .

We know that 𝑦 1 ( π‘₯ ) = c o s 𝑝 ( π‘₯ ) , 𝑦 2 ( π‘₯ ) = s i n 𝑝 ( π‘₯ ) are two linearly independent solutions of (1.3), so the general solution of (1.3) can be expressed as 𝑦 ( π‘₯ ) = 𝑐 1 c o s 𝑝 ( π‘₯ ) + 𝑐 2 s i n 𝑝 ( π‘₯ ) . ( 2 . 1 ) Because of | 𝑝 ( π‘₯ ) ( 0 + ) | = | 𝑝 ( π‘₯ ) ( 1 βˆ’ ) | = ∞ , the function 𝑦 ( π‘₯ ) is two-point oscillatory on 𝐼 , then (1.3) is two-point oscillatory on 𝐼 .

Example 2.3. Let 𝑝 ( π‘₯ ) = βˆ’ l n l n ( 1 / π‘₯ ) , π‘₯ ∈ 𝐼 , then 𝑝 ξ…ž ( π‘₯ ) = 1 / ( π‘₯ l n ( 1 / π‘₯ ) ) , 𝑝 ξ…ž ξ…ž ( π‘₯ ) = ( 1 βˆ’ l n ( 1 / π‘₯ ) ) / ( π‘₯ l n ( 1 / π‘₯ ) ) 2 , 𝑝 ( π‘₯ ) satisfies the condition (1.4), so the following equation: ξ‚΅ 1 𝑦 π‘₯ l n ( 1 / π‘₯ ) ξ…ž ξ‚Ά ξ…ž βˆ’ 2 1 βˆ’ l n ( 1 / π‘₯ ) ( π‘₯ l n ( 1 / π‘₯ ) ) 2 𝑦 ξ…ž + ξ‚΅ 1 ξ‚Ά π‘₯ l n ( 1 / π‘₯ ) 3 𝑦 = 0 ( 2 . 2 ) is two-point oscillatory on 𝐼 .

Example 2.4. Let 𝑝 ( π‘₯ ) = βˆ’ ( 1 βˆ’ 2 π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) πœ€ , π‘₯ ∈ 𝐼 , where πœ€ > 0 . Then, 𝑝 ξ…ž ( 2 ξ€· π‘₯ ) = π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ ( πœ€ + 1 ) , 𝑝 ξ…ž ξ…ž ξ€· ( π‘₯ ) = βˆ’ ( πœ€ + 1 ) ( 1 βˆ’ 2 π‘₯ ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ πœ€ ξ€Ί 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ξ€» ξ€· βˆ’ ( 2 βˆ’ 4 πœ€ ) ( 1 βˆ’ 2 π‘₯ ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ πœ€ + 1 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 2 πœ€ + 2 , ( 2 . 3 ) when π‘₯ β†’ 0 , 𝑝 ( π‘₯ ) β†’ βˆ’ ∞ ; when π‘₯ β†’ 1 , 𝑝 ( π‘₯ ) β†’ βˆ’ ∞ , which satisfies the condition (1.4); 𝑝 ξ…ž ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) , 𝑝 ξ…ž ξ…ž ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) . Substituting 𝑝 ξ…ž ( π‘₯ ) and 𝑝 ξ…ž ξ…ž ( π‘₯ ) into (1.3), we obtain that the following equation:  2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ( π‘₯ βˆ’ π‘₯ 2 ) ( πœ€ + 1 ) 𝑦 ξ…ž ξƒͺ ξ…ž ξ€· βˆ’ 2 ( 2 βˆ’ 4 πœ€ ) ( 1 βˆ’ 2 π‘₯ ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ πœ€ + 1 ξ€· βˆ’ ( πœ€ + 1 ) ( 1 βˆ’ 2 π‘₯ ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ πœ€ ξ€Ί 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ξ€» ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 2 πœ€ + 2 𝑦 ξ…ž +  2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ ( πœ€ + 1 ) ξƒͺ 3 𝑦 = 0 ( 2 . 4 ) is two-point oscillatory on 𝐼 .

3. A New Comparison Theorem

Theorem 3.1. Suppose that the second order differential equations ξ€· 𝑝 1 ( π‘₯ ) 𝑦 ξ…ž ξ€Έ ( π‘₯ ) ξ…ž + π‘Ÿ 1 ( π‘₯ ) 𝑦 ξ…ž ( π‘₯ ) + π‘ž 1 ξ€· 𝑝 ( π‘₯ ) 𝑦 ( π‘₯ ) = 0 , ( 3 . 1 ) 2 ( π‘₯ ) 𝑧 ξ…ž ξ€Έ ( π‘₯ ) ξ…ž + π‘Ÿ 2 ( π‘₯ ) 𝑧 ξ…ž ( π‘₯ ) + π‘ž 2 ( π‘₯ ) 𝑧 ( π‘₯ ) = 0 , ( 3 . 2 ) satisfy the existence and uniqueness theorem on 𝐼 , and one of the following conditions holds: (1)when 0 < 𝑝 2 < 𝑝 1 and π‘Ÿ 1 β‰  π‘Ÿ 2 , π‘ž 2 β‰₯ π‘ž 1 + π‘Ÿ 2 2 4 𝑝 2 + ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ , ( 3 . 3 ) (2)when 0 < 𝑝 1 = 𝑝 2 = 𝑝 and π‘Ÿ 1 β‰  π‘Ÿ 2 , there exists an 𝑛 ∈ ℝ + , which satisfies π‘ž 2 β‰₯ ( 1 + 𝑛 ) π‘ž 1 + ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 + π‘Ÿ 2 2 4 𝑝 , ( 3 . 4 ) then (3.2) has at least one zero point between two consecutive zero point 𝛼 , 𝛽 ( 𝛼 < 𝛽 ) of any nontrivial solution 𝑦 ( π‘₯ ) of (3.1).

Proof. (1) We suppose that 𝑧 ( π‘₯ ) has no zero point on [ 𝛼 , 𝛽 ] when 0 < 𝑝 2 < 𝑝 1 . Without loss of generality, let 𝑦 ( π‘₯ ) β‰₯ 0 , 𝑧 ( π‘₯ ) > 0 , π‘₯ ∈ [ 𝛼 , 𝛽 ] , then we have 𝑑  𝑦 𝑑 π‘₯ 𝑧 ξ€· 𝑧 𝑝 1 𝑦 ξ…ž βˆ’ 𝑦 𝑝 2 𝑧 ξ…ž ξ€Έ ξ‚„ = 𝑦 𝑧  𝑧 ξ€· 𝑝 1 𝑦 ξ…ž ξ€Έ ξ…ž + 𝑝 1 𝑧 ξ…ž 𝑦 ξ…ž ξ€· 𝑝 βˆ’ 𝑦 2 𝑧 ξ…ž ξ€Έ ξ…ž βˆ’ 𝑝 2 𝑦 ξ…ž 𝑧 ξ…ž ξ‚„ + 𝑦 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 𝑧 2 ξ€· 𝑝 1 𝑦 ξ…ž 𝑧 βˆ’ 𝑝 2 𝑦 𝑧 ξ…ž ξ€Έ = 𝑦 𝑧 ξ€Ί 𝑧 ξ€· βˆ’ π‘ž 1 𝑦 βˆ’ π‘Ÿ 1 𝑦 ξ…ž ξ€Έ ξ€· βˆ’ 𝑦 βˆ’ π‘ž 2 𝑧 βˆ’ π‘Ÿ 2 𝑧 ξ…ž ξ€Έ + ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ 𝑦 ξ…ž 𝑧 ξ…ž ξ€» + 𝑦 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 𝑧 2 ξ€· 𝑝 1 𝑦 ξ…ž 𝑧 βˆ’ 𝑝 2 𝑦 𝑧 ξ…ž ξ€Έ = ξ€· π‘ž 2 βˆ’ π‘ž 1 ξ€Έ 𝑦 2 + π‘Ÿ 2 𝑦 2 𝑧 ξ…ž 𝑧 βˆ’ π‘Ÿ 1 𝑦 𝑦 ξ…ž + ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ 𝑦 𝑦 ξ…ž 𝑧 ξ…ž 𝑧 + 𝑝 1 ξ€· 𝑦 ξ…ž ξ€Έ 2 βˆ’ 𝑝 1 𝑦 𝑦 ξ…ž 𝑧 ξ…ž 𝑧 βˆ’ 𝑝 2 𝑦 𝑦 ξ…ž 𝑧 𝑧 ξ…ž 𝑧 2 + 𝑝 2 ξ€· 𝑦 𝑧 ξ…ž ξ€Έ 2 𝑧 2 = ξ€· π‘ž 2 βˆ’ π‘ž 1 ξ€Έ 𝑦 2 + π‘Ÿ 2 𝑦 2 𝑧 ξ…ž 𝑧 βˆ’ π‘Ÿ 1 𝑦 𝑦 ξ…ž + ξ€· 𝑝 1 βˆ’ 𝑝 2 𝑦 ξ€Έ ξ€· ξ…ž ξ€Έ 2 + 𝑝 2 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά 2 = ξ€· π‘ž 2 βˆ’ π‘ž 1 ξ€Έ 𝑦 2 + ξ€· √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξ…ž ξ€Έ 2 βˆ’ ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 𝑦 𝑦 ξ…ž βˆ’ π‘Ÿ 2 𝑦 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά + ξ‚Έ √ 𝑝 2 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά ξ‚Ή 2 +  π‘Ÿ 1 βˆ’ π‘Ÿ 2 2 √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξƒͺ 2 +  π‘Ÿ 2 𝑦 2 √ 𝑝 2 ξƒͺ 2 βˆ’ ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ 𝑦 2 βˆ’ π‘Ÿ 2 2 4 𝑝 2 𝑦 2 =  √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξ…ž βˆ’ π‘Ÿ 1 βˆ’ π‘Ÿ 2 2 √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξƒ­ 2 +  π‘Ÿ 2 𝑦 2 √ 𝑝 2 βˆ’ √ 𝑝 2 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά ξƒ­ 2 +  π‘ž 2 βˆ’ π‘ž 1 βˆ’ π‘Ÿ 2 2 4 𝑝 2 βˆ’ ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ ξƒ­ 𝑦 2 . ( 3 . 5 ) Integrating the above equation from 𝛼 to 𝛽 , we obtain ξ€œ 0 = 𝛽 𝛼  √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξ…ž βˆ’ π‘Ÿ 1 βˆ’ π‘Ÿ 2 2 √ 𝑝 1 βˆ’ 𝑝 2 𝑦 ξƒ­ 2 + ξ€œ 𝑑 π‘₯ 𝛽 𝛼  π‘Ÿ 2 𝑦 2 √ 𝑝 2 βˆ’ √ 𝑝 2 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά ξƒ­ 2 + ξ€œ 𝑑 π‘₯ 𝛽 𝛼  π‘ž 2 βˆ’ π‘ž 1 βˆ’ π‘Ÿ 2 2 4 𝑝 2 βˆ’ ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ ξƒ­ 𝑦 2 𝑑 π‘₯ , ( 3 . 6 ) that is, ξ€œ 0 β‰₯ 𝛽 𝛼  π‘ž 2 βˆ’ π‘ž 1 βˆ’ π‘Ÿ 2 2 4 𝑝 2 βˆ’ ξ€· π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ ξƒ­ 𝑦 2 𝑑 π‘₯ . ( 3 . 7 ) From previous equality and assumption (3.3), we obtain the next equalities: √ 𝑝 1 βˆ’ 𝑝 2 π‘Ÿ 𝑦 β€² = 1 βˆ’ π‘Ÿ 2 2 √ 𝑝 1 βˆ’ 𝑝 2 π‘Ÿ 𝑦 , ( 3 . 8 ) 2 𝑦 2 √ 𝑝 2 = √ 𝑝 2 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά . ( 3 . 9 ) By (3.8), we obtain 𝑦 ξ…ž / 𝑦 = ( π‘Ÿ 1 βˆ’ π‘Ÿ 2 ) / 2 ( 𝑝 1 βˆ’ 𝑝 2 ) , 𝑦 = 𝑒 ∫ ( ( π‘Ÿ 1 βˆ’ π‘Ÿ 2 ) / 2 ( 𝑝 1 βˆ’ 𝑝 2 ) ) 𝑑 π‘₯ . By (3.9), we obtain ( 𝑦 ξ…ž / 𝑦 ) βˆ’ ( 𝑧 ξ…ž / 𝑧 ) = ( π‘Ÿ 2 / 2 𝑝 2 ) . In summary, we obtain 𝑧 = 𝑦 𝑒 βˆ’ ∫ ( π‘Ÿ 2 / 2 𝑝 2 ) 𝑑 π‘₯ , that is, 𝑧 ( 𝛼 ) = 𝑧 ( 𝛽 ) = 0 , which contradicts with the assumption.
(2) We suppose that 𝑧 ( π‘₯ ) has no zero point on [ 𝛼 , 𝛽 ] when 0 < 𝑝 1 = 𝑝 2 = 𝑝 . Without loss of generality, let 𝑦 ( π‘₯ ) β‰₯ 0 , 𝑧 ( π‘₯ ) > 0 , π‘₯ ∈ [ 𝛼 , 𝛽 ] , for all 𝑛 ∈ ℝ + , then 𝑑  𝑦 𝑑 π‘₯ 𝑧 ξ€· 𝑧 𝑝 𝑦 ξ…ž βˆ’ 𝑦 𝑝 𝑧 ξ…ž ξ€Έ + 𝑛 𝑦 𝑝 𝑦 ξ…ž ξ‚„ = 𝑦 𝑧 ξ‚€ 𝑧 ξ€· 𝑝 𝑦 ξ…ž ξ€Έ ξ…ž ξ€· βˆ’ 𝑦 𝑝 𝑧 ξ…ž ξ€Έ ξ…ž  𝑦 + 𝑝 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 𝑧 2 ξ€· 𝑦 ξ…ž 𝑧 βˆ’ 𝑦 𝑧 ξ…ž ξ€Έ ξ€· 𝑦 + 𝑝 ξ…ž ξ€Έ 2 ξ€· + 𝑛 𝑦 βˆ’ π‘ž 1 𝑦 βˆ’ π‘Ÿ 1 𝑦 ξ…ž ξ€Έ = 𝑦 𝑧 ξ€Ί 𝑧 ξ€· βˆ’ π‘ž 1 𝑦 βˆ’ π‘Ÿ 1 𝑦 ξ…ž ξ€Έ ξ€· βˆ’ 𝑦 βˆ’ π‘ž 2 𝑧 βˆ’ π‘Ÿ 2 𝑧 ξ…ž ξ€· 𝑦 ξ€Έ ξ€» + 𝑝 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 ξ€Έ 2 𝑧 2 ξ€· 𝑦 + 𝑝 ξ…ž ξ€Έ 2 ξ€· + 𝑛 𝑦 βˆ’ π‘ž 1 𝑦 βˆ’ π‘Ÿ 1 𝑦 ξ…ž ξ€Έ = ξ‚Έ ξ€· π‘ž 2 βˆ’ ( 1 + 𝑛 ) π‘ž 1 ξ€Έ 𝑦 2 + π‘Ÿ 2 𝑦 2 𝑧 ξ…ž 𝑧 βˆ’ ( 1 + 𝑛 ) π‘Ÿ 1 𝑦 𝑦 ξ…ž ξ‚Ή ξ€· 𝑦 + 𝑝 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 ξ€Έ 2 𝑧 2 ξ€· 𝑦 + 𝑝 ξ…ž ξ€Έ 2 = ξ‚Έ ξ€· π‘ž 2 βˆ’ ( 1 + 𝑛 ) π‘ž 1 ξ€Έ 𝑦 2 βˆ’ ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 𝑦 𝑦 ξ…ž βˆ’ π‘Ÿ 2 𝑦 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ€· 𝑦 ξ‚Ά ξ‚Ή + 𝑝 ξ…ž 𝑧 βˆ’ 𝑧 ξ…ž 𝑦 ξ€Έ 2 𝑧 2 ξ€· 𝑦 + 𝑝 ξ…ž ξ€Έ 2 = ξ€· π‘ž 2 βˆ’ ( 1 + 𝑛 ) π‘ž 1 ξ€Έ 𝑦 2 βˆ’ ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 𝑦 𝑦 ξ…ž βˆ’ π‘Ÿ 2 𝑦 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά ξ‚΅ 𝑦 + 𝑝 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά 2 ξ€· 𝑦 + 𝑝 ξ…ž ξ€Έ 2 + π‘Ÿ 2 2 𝑦 4 𝑝 2 + ξ€· 2 π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 𝑦 4 𝑝 2 βˆ’ ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 𝑦 4 𝑝 2 βˆ’ π‘Ÿ 2 2 𝑦 4 𝑝 2 =  π‘Ÿ 2 𝑦 2 √ 𝑝 βˆ’ √ 𝑝 ξ‚΅ 𝑦 ξ…ž βˆ’ 𝑦 𝑧 ξ…ž 𝑧 ξ‚Ά ξƒ­ 2 +  ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 𝑦 2 √ 𝑝 βˆ’ 𝑦 ξ…ž √ 𝑝 ξƒ­ 2 +  π‘ž 2 βˆ’ ( 1 + 𝑛 ) π‘ž 1 βˆ’ ξ€· 2 π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 + π‘Ÿ 2 2 ξƒ­ 𝑦 4 𝑝 2 . ( 3 . 1 0 ) Integrating the above equation from 𝛼 to 𝛽 , we obtain ξ€œ 0 β‰₯ 𝛽 𝛼  π‘ž 2 βˆ’ ( 1 + 𝑛 ) π‘ž 1 βˆ’ ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 + π‘Ÿ 2 2 ξƒ­ 𝑦 4 𝑝 2 𝑑 π‘₯ . ( 3 . 1 1 ) We can find the contradiction similarly; here, we delete the details. This completes the proof.

When 𝑧 ( π‘₯ ) = 𝑀 ( 𝑧 ( π‘₯ ) ) is a nonlinear term, where 𝑀 ( 𝑧 ) ∢ 𝑅 β†’ 𝑅 is a continuous function and 𝑧 𝑀 ( 𝑧 ) > 0 for 𝑧 β‰  0 , Zhuang and Wu established some comparison theorems if 𝑀 β€² ( 𝑧 ) β‰₯ 𝐾 > 0 holds in [3]. The condition of Corollary 2.2 in [3] is identical with (3.3) when 𝑀 ξ…ž ( 𝑧 ) is smooth and 𝐾 = 1 , but there's no condition about the situation of 𝑝 1 = 𝑝 2 = 𝑝 . We put β€œ 𝑛 𝑦 𝑝 ξ…ž 𝑦 ” added to Picone identity, which solve the problem of the vacuousness of (3.3) when 𝑝 1 = 𝑝 2 = 𝑝 . Then, we obtain (3.4) and establish the integrated comparison theorem of second order damped linear differential equations.

We can easily obtain the following corollaries by Theorem 3.1.

Corollary 3.2. Suppose (3.1), (3.2) satisfy the existence and uniqueness theorem on 𝐼 . If (3.1) is two-point oscillatory on 𝐼 , r 1 β‰  r 2 and satisfies one of the following conditions: (1)when 0 < 𝑝 2 < 𝑝 1 , the following condition is satisfied on 𝐼 , π‘ž 2 β‰₯ π‘ž 1 + π‘Ÿ 2 2 4 𝑝 2 + ξ€· π‘Ÿ 2 βˆ’ π‘Ÿ 1 ξ€Έ 2 4 ξ€· 𝑝 1 βˆ’ 𝑝 2 ξ€Έ , ( 3 . 1 2 ) (2)when 0 < 𝑝 1 = 𝑝 2 = 𝑝 , the following condition is satisfied on 𝐼 , π‘ž 2 β‰₯ ( 1 + 𝑛 ) π‘ž 1 + ξ€· ( 1 + 𝑛 ) π‘Ÿ 1 βˆ’ π‘Ÿ 2 ξ€Έ 2 + π‘Ÿ 2 2 4 𝑝 , ( 3 . 1 3 ) then (3.2) is two-point oscillatory on 𝐼 .

Corollary 3.3. Consider the second order equation (1.3) and the following equation: ξ€· 𝐴 ( π‘₯ ) 𝑦 ξ…ž ξ€Έ ξ…ž + 𝐡 ( π‘₯ ) 𝑦 ξ…ž + 𝐢 ( π‘₯ ) 𝑦 = 0 , ( 3 . 1 4 ) where π‘₯ ∈ 𝐼 , 𝑝 ( π‘₯ ) satisfies condition (1.4), 𝐡 ( π‘₯ ) β‰  βˆ’ 2 𝑝 ξ…ž ξ…ž ( π‘₯ ) . Suppose they satisfy the existence and uniqueness theorem on 𝐼 . When π‘₯ β†’ 0 + and π‘₯ β†’ 1 βˆ’ , if 0 < 𝐴 ( π‘₯ ) < 𝑝 ξ…ž ( π‘₯ ) is satisfied on 𝐼 , and ξ€· 𝑝 𝐢 ( π‘₯ ) β‰₯ ξ…ž ξ€Έ ( π‘₯ ) 3 + 𝐡 ( π‘₯ ) 2 + ξ€· 4 𝐴 ( π‘₯ ) 𝐡 ( π‘₯ ) + 2 𝑝 ξ…ž ξ…ž ξ€Έ ( π‘₯ ) 2 4 ( 𝑝 ξ…ž ( π‘₯ ) βˆ’ 𝐴 ( π‘₯ ) ) , ( 3 . 1 5 ) then (3.14) is two-point oscillatory on 𝐼 .

Remark 3.4. The two-point oscillation of (1.2) is studied by comparison theorem and two-point oscillatory equation in [1]. When 𝑝 1 ( π‘₯ ) = 𝑝 2 ( π‘₯ ) = 1 and π‘Ÿ 1 ( π‘₯ ) = π‘Ÿ 2 ( π‘₯ ) = 0 , Theorem 3.1 reduces to Theorem 2.1 in [1].

As an application of Corollary 3.3, we discuss the two-point oscillation of (1.5). Since Example 2.4 is the known two-point oscillatory equation, that is 𝑝 ( π‘₯ ) = βˆ’ ( 1 βˆ’ 2 π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) πœ€ , π‘₯ ∈ 𝐼 , where πœ€ > 0 , 𝑝 ξ…ž ( π‘₯ ) ∼ ( π‘₯ βˆ’ π‘₯ 2 ) βˆ’ ( πœ€ + 1 ) as π‘₯ β†’ 0 + or π‘₯ β†’ 1 βˆ’ ; 𝑝 ξ…ž ξ…ž ( π‘₯ ) ∼ ( π‘₯ βˆ’ π‘₯ 2 ) βˆ’ ( πœ€ + 2 ) as π‘₯ β†’ 0 + or π‘₯ β†’ 1 βˆ’ . For (1.5), 𝐴 ( π‘₯ ) = 1 , 𝐡 ( π‘₯ ) = 𝑓 ( π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) 𝛼 , 𝐢 ( π‘₯ ) = 𝑔 ( π‘₯ ) / ( π‘₯ βˆ’ π‘₯ 2 ) 𝛽 . Because of 𝑓 ( π‘₯ ) > 0 , 𝑓 ( π‘₯ ) ∈ 𝐢 ( 𝐼 ) , there exists 𝑀 > 0 such that 𝑓 ( π‘₯ ) < 𝑀 for all π‘₯ ∈ 𝐼 . Therefore, 𝐡 ( π‘₯ ) = 𝑓 ( π‘₯ ) ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 𝛼 ∼ ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 𝛼 𝐢 , a s π‘₯ ⟢ 0 + o r π‘₯ ⟢ 1 βˆ’ , ( π‘₯ ) = 𝑔 ( π‘₯ ) ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ 𝛽 ∼ ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 𝛽 , a s π‘₯ ⟢ 0 + o r π‘₯ ⟢ 1 βˆ’ , ( 3 . 1 6 ) Thus, when π‘₯ β†’ 0 + and π‘₯ β†’ 1 βˆ’ , 𝑝 ξ…ž ( π‘₯ ) β†’ + ∞ , 𝐴 ( π‘₯ ) ≑ 1 , βˆ’ 2 𝑝 ξ…ž ξ…ž ( π‘₯ ) β‰  𝐡 ( π‘₯ ) , ξ€· 𝐡 ( π‘₯ ) + 2 𝑝 ξ…ž ξ…ž ξ€Έ ( π‘₯ ) 2 4 ( 𝑝 ξ…ž ∼  ξ€· ( π‘₯ ) βˆ’ 𝐴 ( π‘₯ ) ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 𝛼 + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ πœ€ βˆ’ 2 ξ‚„ 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( πœ€ + 1 ) ∼ ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 3 + πœ€ ) + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 2 𝛼 βˆ’ πœ€ βˆ’ 1 ) + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 𝛼 + 1 ) , a s π‘₯ β†’ 0 + o r π‘₯ β†’ 1 βˆ’ , 𝐡 ( π‘₯ ) 2 ∼ ξ€· 4 𝐴 ( π‘₯ ) π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 2 𝛼 ξ€· 𝑝 , a s π‘₯ ⟢ 0 + o r π‘₯ ⟢ 1 βˆ’ , ξ…ž ξ€Έ 3 =  2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ + πœ€ ( 1 βˆ’ 2 π‘₯ ) 2 ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ ( πœ€ + 1 ) ξƒ­ 3 ∼ ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 3 ( πœ€ + 1 ) , a s π‘₯ ⟢ 0 + o r π‘₯ ⟢ 1 βˆ’ . ( 3 . 1 7 ) By (3.17), the following condition need to be satisfied if (3.15) holds, ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 𝛽 β‰₯ ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 3 ( πœ€ + 1 ) + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ 2 𝛼 + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 3 + πœ€ ) + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 2 𝛼 βˆ’ πœ€ βˆ’ 1 ) + ξ€· π‘₯ βˆ’ π‘₯ 2 ξ€Έ βˆ’ ( 𝛼 + 1 ) , a s π‘₯ ⟢ 0 + o r π‘₯ ⟢ 1 βˆ’ , ( 3 . 1 8 ) that is, 𝛽 > m a x { 3 ( πœ€ + 1 ) , 2 𝛼 , 𝛼 + 1 } . ( 3 . 1 9 ) In summary, let πœ€ β†’ 0 , then, when 0 < 𝛼 < 3 / 2 , 3 + 3 πœ€ > m a x { 2 𝛼 , 𝛼 + 1 } , condition (3.19) holds with 𝛽 > 3 , (1.5) is two-point oscillatory on 𝐼 in this case,when 𝛼 > 3 / 2 , 2 𝛼 > m a x { 3 + 3 πœ€ , 𝛼 + 1 } , condition (3.19) holds with 𝛽 > 2 𝛼 , (1.5) is two-point oscillatory on 𝐼 in this case.

Acknowledgments

This work was supported by the National Nature Science Foundation of China (10801089, 11171178) and the National Nature Science Foundation of Shandong Province (ZR2009AQ010).