代做Network Theory Coursework II – Academic Year 2023/24帮做Matlab编程

日期：2024-03-25 04:07

Network Theory

Coursework II - Academic Year 2023/24

The deployment of edge computing infrastructure requires a careful placement of the edge servers in terms of their locations and the associated traffic demand (i.e., the spatial demand of the traffic). In that respect, the overall aim is to improve application latency (this is for applications that require edge cloud support such as cached content and/or offloading of computationally heavy procedures from the mobile device to the edge cloud) and reduce overall data transfer load in emerging demanding mobile applications. In other words, we require edge cloud locations that minimize the latency for users to connect to them whilst at the same time the cost of opening such facilities should also be considered. Within that context, deploying edge servers introduces pragmatic complexities. For example, network operators need to consider the placement of physical edge platform components, i.e., the edge cloud servers. Consider that we have m mobile users and n < m locations to open an edge cloud with the path distance (i.e., routing cost) between user i and candidate edge cloud location j is denoted as dij . Furthermore, the cost of opening the edge cloud j is cj . This cost can be considered as a fixed cost for the edge cloud, and it could be considered as an operational cost (i.e., energy consumption and management of the side) if the provider has already acquired the location or a deployment cost (i.e., capital expenditure) if the provider wants to operate/open an edge cloud on that location.

I.        Provide a binary linear mathematical programming formulation that minimize the number of edge clouds to be opened whilst at the same time we are minimizing the overall routing cost of connecting the mobile users to the edge clouds (as described in the aforementioned cost  dij ). Note that this requirement entails a bi- objective  optimization  problem.  Furthermore,  consider  the  requirement  that  a mobile user can connect to only one edge cloud to access services. Explain your formulation.

II.       Implement the above 0/1 integer linear optimization problem in MATLAB and using the intlinprog(.) function find the optimal allocation of mobile users and edge cloud locations opening for the case of n= 5 mobile users and m = 3 candidate edge cloud locations with the following parameters. The distance d ij matrix between mobile user (i) and edge cloud (j) locations is as follows,

d = [10 15 20;

12 18 25;

14 20 22;

11 16 21;

13 19 24];

The cost of opening at specific locations: c = [50; 60; 70]. Explain your MATLAB code in detail.

III.      Consider the case where each mobile user request has a demand of vi (i = 1, . . , m) Virtual Machines (VMs) and that the capacity per edge cloud in terms of available VMs is vj (j  = 1, . . n). Explain how we should amend (augment) the formulation so that we consider this requirement.