I'm having issues for loading libraries:

Step1:

library("oligo",lib.loc="~/R_LIB")
library("pd.huex.1.0.st.v2",lib.loc="~/R_LIB")

Error:

Loading required package: IRanges

Error: package or namespace load failed for ‘IRanges’:
 Functions found when exporting methods from the namespace ‘IRanges’ which are not S4 generic: ‘endoapply’, ‘mendoapply’
Error: package ‘IRanges’ could not be loaded

Step2:

R CMD INSTALL IRanges_2.14.10.tar.gz -l ~/R_LIB
Error : package ‘BiocGenerics’ 0.24.0 was found, but >= 0.25.3 is required by ‘IRanges’
ERROR: lazy loading failed for package ‘IRanges’
* removing ‘../..path_path....../R_LIB/IRanges’

If I upgrade BiocGenerics I'm afraid I'll have to change/upgrade many other packages, too. Is there a workaround of this which can minimize installations? 

I tried to look for IRanges_2.13.X but in vain. I've Version: 2.12.0 Depends: R (>= 3.1.0), methods, utils, stats, BiocGenerics (>= 0.23.3),
        S4Vectors (>= 0.15.5)

Bioconductor version '3.6'

SessionInfo()
R version 3.4.2
other attached packages:
[1] S4Vectors_0.18.2    BiocGenerics_0.24.0
loaded via a namespace (and not attached):
[1] zlibbioc_1.24.0 compiler_3.4.2  affyio_1.48.0   Biobase_2.38.0 

Thanks.

You should never use install.packages directly, but instead should use biocLite. So you could do

library(BiocInstaller)
biocValid()

And then follow the instructions that are given. But do note that you are using an outdated version of R and Bioconductor, and we only offer support for the current versions. If you want help, install R-3.5.0 and then install all the packages you want using biocLite. Unfortunately there isn't a 'workaround of this that can minimize installations' because the whole idea for R and Bioconductor is that we are continually improving and adding things. If you want to stick with old versions, that's completely at your discretion, but we cannot support anything but the current release versions.