i am not too use to eview but i think you can google search it

These restrictions help to identify the structural errors and must have an economic meaning. The simplest example is to use the Cholesky decomposition, in this case a contemporary order is assumed.

For example , if you have a VAR model for variables X, Y and Z and use a Cholesky ordering  X Y Z, you are assuming  (in terms of the impulse response functions) that shocks in Y and Z do not affect X contemporaneously, and that shocks in Z do not affect Y contemporaneously. When estimating a VAR model in EViews,  the Cholesky ordering is given in the “Impulse definition” option of the window “impulse” of VAR object.

You can also use short-run and long-run restrictions using selecting “Proc/Estimate Structural Factorization” from the VAR window menu. The syntaxes of these restrictions are explained in the section “Structural (Identified) VARs” of Chapter 36 (Vector Autoregression and Error Correction Models) in the EViews 8 Users Guide II.

Suppose you are working with 3 variables X, Y and Z. If you are getting two cointegration vectors then the first vector will be (1 0 beta13)  and second vector will be (0 1 beta23).  It simply means that there exists two equilibrium relationships; in first X depends upon Z and in second Y depends upon Z. Now suppose your economic theory says X depends upon Y and Y must depend upon Z. In such case you may impose restrictions betaa11=1, beta13=0, beta21=0, and b22=1.   Now the new vectors you will get (1 beta12 0) and (0 1 beta23). So you will get the estimates of two desired equilibrium relationships that represents the structure of economy.

Please I need a better understanding of VAR. I am working on "The effect of International Oil Price on inflation and exchange rate" for my thesis. I am thinking of using VAR then interpret using IRF and FEVD. Any suggestions please...?

I just made a tutorial about this:

https://youtu.be/9pLKAdhDpOo

a VAR is a reduced form (RF) model. Structural relationships cannot be deduced from VAR coefficients without imposing identifying restrictions on what structure is possible, A simple example is to assert that a causal chain exists in the structure. This sets many structural coefficients to zero, allowing the others to be deduced from the RF VAR. Causal chains are unlikely in many contexts so a variety of other bases for restrictions have been proposed. For an introduction by a leading author you could look at https://cadmus.eui.eu/bitstream/handle/1814/19354/ECO_2011_30.pdf