Research Article | Open Access
Esmaeil Siahlooei, Seyed Abolfazl Shahzadeh Fazeli, "Two Iterative Methods for Solving Linear Interval Systems", Applied Computational Intelligence and Soft Computing, vol. 2018, Article ID 2797038, 13 pages, 2018. https://doi.org/10.1155/2018/2797038
Two Iterative Methods for Solving Linear Interval Systems
Conjugate gradient is an iterative method that solves a linear system , where is a positive definite matrix. We present this new iterative method for solving linear interval systems , where is a diagonally dominant interval matrix, as defined in this paper. Our method is based on conjugate gradient algorithm in the context view of interval numbers. Numerical experiments show that the new interval modified conjugate gradient method minimizes the norm of the difference of and at every step while the norm is sufficiently small. In addition, we present another iterative method that solves , where is a diagonally dominant interval matrix. This method, using the idea of steepest descent, finds exact solution for linear interval systems, where ; we present a proof that indicates that this iterative method is convergent. Also, our numerical experiments illustrate the efficiency of the proposed methods.
Solving system of linear equations is a well-known problem in linear algebra. Many practical problems are modeled as system of linear equations. These problems have been studied by many scientists and several methods have been proposed to solve them [1, 2]. But, in everyday life, measuring instruments just estimate values, and usually measured values are not accurate. Sometimes for more accuracy, intervals are used to represent the actual values. For example, the length of a metal rod is estimated between and centimeters, or temperature measured by a thermometer is between and Celsius.
There are many types of uncertainty and there are many different mathematical systems that calculate uncertainties , e.g., rough sets theory , fuzzy numbers , probability theory , interval valued numbers , and dual fuzzy numbers . In this paper, uncertainty is considered as interval numbers. Some articles used this model to solve their problems [9, 10].
Using interval numbers in algebra was initially developed in the mid-1960s. In 1966, Moore presented his book on interval analysis . Then Hansen offered a solution on interval linear algebraic equations . Then many authors published their methods for solving linear interval systems, such as Neumaier, Abolmasoumi and Alavi, Nirmala, and Ganesan [13–15]. Nowadays, interval analysis methods have been applied to engineering problems, such as dynamic response analysis [16, 17], geotechnical structures , and control systems .
In this paper, we introduce two new iterative methods for solving a linear interval system of equations that is a linear system involving uncertain coefficients appearing as interval numbers. The solution of this system is an interval vector.
The present paper is organized as follows. The basic definitions related to interval numbers and linear interval systems are discussed in Section 3. In Section 3.1, we introduce two additional interval operations. These are inverse operations of “−” and “+.” Section 3.2 recalls some properties of interval numbers and interval arithmetic. Also, it introduces some new properties on introduced operations. Next, we review conjugate gradient method and then present a new method for solving linear interval systems based on conjugate gradient method. Another new iterative method using steepest descent idea is proposed in Section 3.4. Section 4 shows the experimental results of the proposed methods and discusses the accuracy and efficiency of the new methods.
2. Interval Arithmetic
Interval numbers and arithmetic are explained in [7, 11, 13, 20]. We review main definition here. Given , where , the real bounded set is called a proper interval. The set of all proper intervals on is denoted by . In this paper, all elements in are shown with a hat, i.e., .
The infimum and supremum of are and , respectively.
The magnitude is defined to be the maximum value of for all . ThusThe width and midpoint of are denoted by and , respectively, and defined asAn interval number is defined as a subset of interval number when and and denoted by . Equality held when .
Each real number can be viewed as a special interval number . The interval simply can be denoted by without confusion.
For each binary operation which is defined on , a binary operation is defined on asfor . Note that we will use operator for both interval and real numbers when there is no confusion.
Therefore, basic interval arithmetic is defined in the following. Let ; thenwhereNote that there is a simple alternative way to formulate the interval multiplication.
An interval vector is a vector whose elements are interval numbers. Similarly, an interval matrix is a matrix whose elements are interval numbers.
Consider , , , and . We say if each element of vector belongs to corresponding element of and if each element of matrix belongs to the corresponding element of .
Norm of interval vector is defined in  asWe define a square interval matrix as diagonally magnitude dominant or for simplicity diagonally dominant if
Proposition 1. Interval matrix is diagonally dominant if and only if
Proof. Equation (13) directly follows (12); just set and when .
If (13) holds then andFinally,General matrix form of linear interval systems can be written as follows:where is an interval matrix, is an interval vector solution, and is an interval vector.
To solve a linear interval system, there exist some different approaches [22–24]. Four of the well-known approaches are as follows:(1)Find an interval vector such that, for all , and exist such that (united solution set).(2)Find an interval vector such that, for all , for all , exists such that (tolerable solution set). Equivalently,(3)Find an interval vector such that, for all , for all , exists, such that (controllable solution set). Equivalently,(4)Find an interval vector such that multiplication is equal to (exact solution set). United solutions are studied in [13, 25, 26]. In [22, 27] methods have been published to find tolerable solutions. Also, some methods have been developed to obtain controllable solutions [14, 28] and exact solutions [12, 14, 29–31] of interval linear systems.
Our proposed method calculates a solution for in the sense of the exact solution, where is a diagonally dominant interval matrix.
3. Materials and Methods
3.1. Inverse Operations for “+” and “”
Suppose we are given an equation , where is a known interval number, and is an unknown interval number, how can one find ?
Obviously is the solution to this equation. However, which is not equal to unless . If the operation “−” is inverse of the operation “+,” then must equal . But and unless .
As it can be observed, the operator “−” is not inverse of the operation “+.” Similarly, it can be observed that the operation “” is not inverse of the operation “.” Therefore, we propose the inverse of operations “+” and “.”
Before introducing inverse operations for “+” and “,” we need another general type of intervals that is defined in [32, 33]. A general closed interval is identified by two real numbers and is defined asThe general interval number space over real numbers is denoted by . Obviously, . All definitions (supremum, infimum, and magnitude) and operations in are defined the same as but computation over is more complicated than . The arithmetic over is well discussed in [20, 32, 33]. For example, as in (7), or as in (8). The width of a general interval number can be negative; for example, . Also, according to definition of subset, .
In this paper, by general interval, we mean general closed intervals.
For , the operation “” is defined in [20, 33] asFor , we define the operation “” asNote that this definition can handle many cases where belongs to , and just in three cases compute undefined or unbounded solution. This property lets us design algorithms with high stability and consistent when working with interval values that contain .
In the following two theorems, we show that operations “” and “” are inverse of operations “+” and “,” respectively .
Theorem 2. Given , then the interval equation has a unique solution , where . The equation has a unique solution in if ; otherwise the equation has no solution in but has exactly one in .
Proof. That is, , which solves . Similarly, ; therefore, we have , where .
Therefore if and , then can be found with this inverse operation uniquely.
Theorem 3. Given , then the interval equation has a solution , where . The equation has a solution in if ; otherwise the equation has no solution in but has in .
Remark. Solution is not usually unique; e.g., the interval equation has solutions . Actually, this equation has unlimited solutions. All intervals and where and can be solution of this equation.
In cases that is not unique, is considered as the longest interval as possible. In the above example, must be .
Proof. Suppose . Before starting the proof, note that multiplication result, , depends on the sign of and . Table 1 shows sign of related to sign of and and also shows formulas of multiplication result.
Suppose that, given , , and and . Now, from Table 1, it is observed that the following must occur: and . ThenThis means in this case is uniquely calculated by inverse operations of “.” For other cases except cases with condition , one can provide similar proofs.
For cases with condition , there is an ambiguity. From Table 1, there are three cases in column . They have the same conditions . The solution can be in forms , , and , respectively. The solution should be in form of , because it is the longest possible interval solution and includes both of the others. Now, find and . If with condition is the solution of , then . It follows thatandThen,If , then , and . This contradicts with . HenceSimilarly,and therefore, we have .
In addition, special case should be considered separately. If and , no exists such that . If , then is free and expression is consistent for all . At this point is considered as .
3.2. Some Properties of the Interval Arithmetic Operations
Proposition 4. Consider ; then
Proof. Let ; then and , so
Proposition 5. For we haveand equality holds when or .
Proof. Let and then . SoObviously equality holds when or , where , which implies property (36).
Proposition 6. For we have
Proof. and using definition of we haveDefine and ; then and . So,and hence (40) holds.
3.3. Modified Conjugate Gradient Method for Linear Interval Systems
Conjugate gradient method is one of the most useful techniques in solving iterative methods for solving linear system of equations, whose matrix is symmetric and positive definite. The conjugate gradient method was proposed by Hestenes and Stiefel in 1952 as an iterative method for solving linear systems with positive definite coefficient matrices . The conjugate gradient algorithm for linear system of equations can be briefly described as follows.
Consider system with solution . Start with an initial estimate of . At step use to new estimate of which is closer to . At each step residual is computed and used as a measure of the goodness of the estimate . The algorithm is detailed below for solving , where A is a symmetric, positive definite matrix.
We generalize this method for linear interval systems when the interval matrix is symmetric, for all , and is an interval vector. There are articles describing and analyzing conjugate gradient method such as [37–39]. These articles illustrated which operator was used in original role, or acted in inverse operator. The new algorithm is obtained from original conjugate gradient method by modifying and replacing all operations with related interval operations; also all operators acting in the role of inverse operators are replaced by inverse of interval operators. In addition, we use interval vector norm instead of real vector norm.
In the first modification on conjugate gradient, we change some operation “−,” with inverse operation “” whenever needed. In Step 2 of Algorithm 1, “−” is applied for computing difference of and . In Step 7 operation “−” is used, but at proof of conjugate gradient method in , is replaced with expression . So, the operation “+” is summed over and . This means “−” in Step 7 is used for inverse of “+.” Therefore we use the operation “” instead of “−” in steps 2 and 7.
The second modification is similar to the first modification. We change operation “” with inverse operation “.” In steps 5 and 8 of Algorithm 1 operation “” is used as inverse of multiplication in proof of conjugate gradient .
The third modification, , is used instead of . In real linear algebra, expression equals square of of , and value of is related to amount of error. We replace this with that has the same meaning. The interval modified conjugate gradient method is the result of these modifications and is shown in Algorithm 2.
3.4. Interval Steepest Descent Method for Linear Interval Systems
First, we define derivative on interval functions. In differential algebra, the derivative of with respect to measures the sensitivity to change value of with respect to . In mathematical termsAccording to this, the derivative of an interval function with respect to is denoted as or or and is defined by the following formula:Let and be a function. The partial derivative of a multivariable interval function is a derivative with respect to one variable when all other variables are fixed. The partial derivative with respect to is denoted as and is defined as Immediately from definitions (44) and (45) and the properties discussed, one can obtain the following properties of derivative and partial derivative of interval functions:Now consider the linear interval system when is a symmetric interval matrix, and is a known interval vector. Let be an interval functionThe gradient of with respect to is the vector of partial derivatives asLet and . Thenso, for any ()And,so, for all ()Note that . Thenand then, from (52)Assume is the solution of linear interval system ; thenSo,Now, consider and , where element of is and other elements of are zero . Note that indicates the opposite of direction . ThenThe derivative of with respect to isIf equals zero, then one can obtain asNow