While it can be debated the relative benefits and costs of mutation, the existence of mutations is not so much a consequence of evolution as it is a consequence of mechanics. The cell cycle is a short, nasty process and mitoses must take place in the S phase of a cell, which itself lasts only a couple hours (if that). But the length of one strand of cellular DNA is somewhere between 2 - 3 meters for humans, and must be straitened out and copied within the confines of the cell's nucleus in the period of an hour. And, this DNA is very, very thin, only a few molecules wide.
So, analagy: the developed northern half of New Jersey is the size of a single cell. Take a shoelace that is about the length of the width across the United States (New York to LA) and roll up most of it around some golf balls (twice around each, specifically). Give one end to a gas-guzzling SUV (this takes energy), and the other to a hummer. Expose about a mile of it and drive 100mph down the New Jersey Turnpike (our site of DNA replication) going two opposite directions. With the ball of yarn next to you, weave out the sequence you see on the shoelace as it races past you while you dodge the traffic of other molecules. It's semi-conservative replication, so two of these are happening at the same time. And we have some 24 chromosomes to do. You have an hour, and for simplicity we'll leave out the bit about Okazaki fragments.
This is where mutations come from.
‘ ‘ — — ‘
™ Version 4.2