Why do majority of the PE fund value the company based on EV/EBITA ratio ?

Unsure if you meant EBITDA or EBITA, as there is a subtle difference in the two. EBITDA shows earnings before depreciation and amortisation, the other takes depreciation into account but not amortisation.

Why do we use an EV/EBITDA or EV/EBITA or EV/EBIT multple?

Enterprise Value (EV) tells you the total value of the business i.e. how much in total you need to pay off the equity and debt owners, if you acquire the company at debt free cash free level

Why is the EV divided by EBITDA or EBITA or EBIT?

EBITDA is used because it gives a true picture of company's earning without the influence of capital structure

EBIT is used because Capex are not free or paid by an angel. In capital intensive industries, you will likely come across business where majority of their EBITDA is locked in depreciation and amortisation - so how would you value that as you are not taking their true earnings

EBIT takes into account the depreciation and amortisation, which therefore gives a more closer view of the business' earnings in those kind of capital intensive industries 

EBITDA is slightly disliked in that scenario because it excludes the often sizable capex companies make and it hides how much cash they are actually using to finance their operations.

Though before any EV multiples, every PE house will definitely value the business from a LBO (Leveraged buy-out) perspective. I have never come across a PE house that will not do a LBO modelling and just value the business at an EV/EBITDA level!

They would do a LBO valuation purely because every PE model is based on the basic theory of leveraging the growth through sourcing of funds largely by debt

PE funds will be investing in startup companies. Since those startups are in early stages,they may not be generating profits, and they require huge investments. So in the immediate 2-3 years,PE funds know these startups wont make profits and hence no point in looking at PE multiples. Some companies going forward,once after the initial investment phase,startups may turn to make operating profits,so its better to go with EV/EBITDA. With this they can see whether a company is able to generate profit operationally. So,over the long term,if the debt is brought down by giving fresh money to those companies,PE funds can make those companies generate profits.

PE Funds have a 3-7 year timeline on most investments between the time of acquisition and the time of exit. From an investment IRR perspective, the goal is to buy low and sell high. The idea is to acquire an investment at a lower EV/EBITDA multiple, add EBITDA to the firm over the 3-7 year holding period (either through acquiring EBITDA through additional acquisitions at a lower multiple than investment EBITDA multiple or implementing strategic or financial engineering to improve the EBITDA organically) and then selling the combined higher EBITDA at a higher exit multiple than acquisition. It's a quick way to show value-add from a PE point of view.

Some of the points mentioned below regarding taking out the impact of capital structure by looking at the EBITDA versus EBIT or NI are also correct. You will not be using P/E because a Private Equity by definition would be investing in a private entity (Public firms EV/EBITDA is used after applying suitable discounts based on several factors involved). 

EV/EBITA doesn't discriminate between equity holders or debt holders, also accounting for the entire firm which provides the real value of firm ignoring the profitability to buy a stake.

PE funds are more interested in looking at Enterprise value multiples rather than enterprise net earnings. Because net earnings leads to bias due to differences in accounting policies and capital structure. Moreover, some firms may charge depreciation on accelerated basis which leads to high depreciation cost in initial years.

In addition, some firm may have high debt in their capital structure which increases the interest cost and depress the net earnings. EBITA discards such difficulties and used as a proxy for cash flow as it add depreciation, the non cash expenditures in order to arrived at a fair value.

EV/EBITDA ratio is a good multiple considering the fact that it encompasses the debt coponent unlike other multiples such as price to earnings.

I think that's because P/E is an inferior valuation criterion vs. EV/EBITDA:

Companies' net income can be temporarily influenced by non-core activities. Furthermore, net income can also be generated from non-cash actvities such as revaluation gains. EBITDA displays the operational profitability of the company and its non-cash components are tend to be less compared to net income.

Besides, owners' credibility may increase or decrease the debt costs of a company,  thereby influencing the bottom line, regardless of the operational performance...

Hence, as a long term valuation criterion, EV/EBITDA is more reliable than P/E.