3GPP is in the process of enumerating functionality for Access Network Discovery and Selection Function (ANDSF). The ANDSF includes data management and control functionality which are essential for providing network discovery and selection function to the User Equipment (UE, MS, mobile station) w.r.t. operators’ policy. A lack of quantitative measures of ANDSF benefits was identified so our task is to identify these benefits. The aim of the thesis is to present ideas for how to obtain such measures and we develop a tool which implements (some of) these ideas and finally we have examples of possible results which will support such a discussion with numbers. This report also discusses different number of use case scenarios (such as ANDSF and USER) which we implement in our simulation tool. In our simulation model we use to impose preferred strategy, reduced delays and losses, and system utilization efficiency. One more thing likes to clarify that the numbers which we use in our model are just examples. It can be any number but they will give us the same behavior of learning. Providing real gains is outside the scope of our study.

The aim of the thesis is to identify quantitative measures of ANDSF benefits. Our contribution is to develop ideas for how to obtain such measures in order to identify ANDSF benefits. We developed a simulation tool which implements some of these ideas. In simulation tool we impose preferred strategies, reduced delays and losses and work on system utilization efficiency in order to identify ANDSF benefits. In our simulation model we use different implementation terminologies like network traffic, access networks, and preferences w.r.t. quality as well as w.r.t. strategies. We have two main scenarios ANDSF and USER; we also perform extensive comparative study in order to evaluate analysis of gains from ANDSF. We have examples of possible results and our numbers are just examples because providing real gains is outside our study. We evaluate our scenarios w.r.t. three main aspects; how well the strategies are fulfilled? How many attempts are required for users? and compare blocking probabilities. The overall conclusion of the thesis is to find out the reasons that why ANDSF is a better choice for network selection as compare to users themselves. Our results shows that ANDSF is a better option for network selection because it helps users towards use their service efficiently as well as avoid blocking; even in real environment the ANDSF becomes aware of less congested access networks in the area the UE is located, and provides instructions to the UE to move the existing IP flows to the less congested access network.