الفهرس Only 14 pages are availabe for public view
 |  | from | 138 |  |  |
| | | from | 138 | | |
Abstract
Resources Allocation In Distributed Processing Systems Using
AI / ()R Techniques
This thesis aims to provide a new approach to allocate resources in
the Distributed Systems (DS), focussing on minimizing the traffic across
the system by exploiting data distribution scheme and allocating the
links capacities of its communication network.
A Decision Support System (DSS) has been developed that helps the
decision makers, in rationale way, to distribute the system data-base and
then, up on the deduced traffic pattern, assigns optimally the links
capacities of the communication network under the problem constraints.
The proposed system offers the following capabilities: .
• It establishes and builds a repository of data for the characteristics
and attributes of the DS under study allowing dynamic adaptation to any
changes in its data such as adding, deleting and/or changing data of
nodes and/or system datasets .
• It builds a starting network topology for the OS being examined,
automatically using a heuristic approach, and provides a complete
graphical user interface (GUI) to enable, the user to change the topology,
interactively and graphically, and hence can in a user-friendly manner
keep variations of the OS network topology for the analysis and decision
making process.
• It provides the capability to have a given network traffic pattern or use
proposed data distribution assessment fuzzy model that can reason how
to locate the distributed system data-base, and hence presents a deduced
t/”(~flic pattern for the DS. It shows, in a GUI, how much data
distribution affects the OS traffic.
• It applies a shortest path algorithm using DP technique to determine
the network traffic flow routs (flow assignment process), based on
deterministic and shortest path routing strategy. It shows in a G U I the
network topology and enable the user to indicate any node-pair and then