Wei Chen, Professor, Dept of Computer Science, TSU

1
INTRODUTION
Wei Chen, Professor, Dept of Computer Science,
TSU

• TOPICS
• Communication Protocols for Sensor Networks
• Parallel/Distributed Processing
• Mobile Ad-Hoc Computing
• - Communication on Ad Hoc Radio Networks
• - Routing Algorithms
• - Autonomous Mobile Robots
• DNA Computing

2
processing
Using a number of processors to finish one task
Speeding up the processing by distributing the
subtasks to the processors Communication time
between the processors is small in a parallel
computer.

• Parallel computing on distributed systems
• Computing is held on a set of distributed
  devices networked PCs and workstations, mobile
  wireless computers, .
• Computing is speeded up by distributing the
  subtasks to the devices.
• Communication time between the devices is very
  large comparing with the computation time.

3
Why parallel/distributed computing?

• Problems with large computing complexity
• Computing hard problems (NP-complete problems)
  exponential computing time.
• Problems with large scale of input size quantum
  chemistry, statistic mechanics, relative physics,
  universal physics, fluid mechanics, biology,
  genetics engineering,
• For example, it costs 1 hour using the current
  computer to simulate the procedure of 1 second
  reaction of protein molecule and water molecule.
  It costs

4
Classification of parallel computers
Processors share a common memory
Processors use distributed memory
Complete connection type
Mash connection type
Hypercube connection type
5
Parallel computation model for parallel algorithm
design

• PRAM(Parallel Random Access Machine)
• PRAM consists of a number of RAM (Random Access
  Machine)
• and a shared memory. Each RAM has a unique
  processor number.
• Processors act synchronously.
• Processor execute the same program.
• (According to the condition fork based on
  processor numbers, it is
• possible to executed different operations.)
• Data communication between processors (RAMs)
• are held through the shared common memory.
• Each processor can write data to and
• read data from one memory cell
• in O(1) time.

6
Our research and results on parallel/distributed
computing

• Optimal parallel algorithms for fundamental
  problems on PRAM and other parallel computational
  models
• Sorting problems, Convex hull problems,
  Shortest path problem,
• Visibility problems, Matrix problems, Envelop
  problems, .
• Hardware design of parallel computers using
  hardware design language.
• One example is the design and implementation of
  PRAM using VHDL.
• Efficient parallel computing on distributed
  computers using PVM/MPI.
• 3-SAT problem, Matrix multiplication, bucket
  sort, .

7
What Is Distribute Systems?
Enslows Definition Distributed System
Distributed hardware Distributed control
Distributed data
Distributed
Physically shared/distributed memory and
logically shared/distributed memory
8
What Is Distribute Systems?
Enslows Definition Distributed System
Distributed hardware Distributed control
Distributed data
Distributed
Physically shared/distributed memory and
logically shared/distributed memory
9
Distributed Computational models

• Processes never mind in which devices they are.
  
• Communicating links communication channels

• Our research distributed system software
• Cooperating the actions of computers.
• Supporting system resources (hard ware and
  software) sharing.
• Supporting data sharing.

10
Communication on Ad Hoc Networks

• What is a Ad Hoc radio network?
• The network consists of a collection of
  transmitter-receiver devises (referred to as
  nodes)
• Each devise s has a transmission range and any
  other device t with this range can directly (i.e.
  by one hop) receive messages from s.
  Communication between two devises that are not
  with in their respective range can be achieved by
  multi-hop transmissions.
• Communication is structured in to synchronous
  time-slots (rounds), a paradigm commonly adopted
  in the practical design of protocols. In every
  round each device acts either as a transmitter or
  as a receiver.
• Knowledge of global topology of the network is
  very limited to each device (especially in a
  mobile network).

11
How to model a Ad Hoc radio network? A Ad Hoc
radio network is modeled by a directed graph
G(V,E), where V is a set of the devices (nodes)
in the radio network, and a directed edge (u,v)
belongs to E iff v can be reached from u.
Example A 6-Hop radio network

• What are fundamental communication tasks?
• Radio Broadcasting (RB) transmitting a message
  from one source node to all other nodes.
• Acknowledged Radio Broadcasting (ARB) achieving
  RB and informing the source about the finish of
  RB.
• Multi broadcasting transmitting a message from
  node source node to some selected nodes.
• Gossiping broadcasting from all nodes to all
  nodes.

12
Autonomous Mobile Robots

• What are autonomous mobile robots in our
  research?
• A set of robots form a distributed system.
• Each robot is viewed as a point and has
  computing capability. It is equipped with sensors
  that let it observes other robots in a local
  range.
• The system is completely decentralized, i.e.,
  the robots are completely autonomous and no
  central control is used.
• Robots have no any global knowledge, such as the
  number of robots, and topology of the system.
• Robots execute a sequence of circles
  Wait-look-Compute-Move.

13
How to model an autonomous robot system? An
autonomous robot system is modeled by a geometric
graph G(V,E), where V is a set of the robots
(nodes) and a directed edge (u,v) belongs to E
iff v locates in the us visible range.
Furthermore, d(u,v) is used to denote the
distance between u and v.
Example Autonomous robots of each with a same
visible range.

• What are fundamental interaction primitives?
• Gathering problem gathering the robots to a
  point autonomously, which is used to coordinate a
  collection of autonomous mobile robots.
• Formation problem distributing the robots to a
  formation autonomously such that the formation
  satisfies some special topology, distribution
  and other properties given in advance.

14
What is our research and results?

• Random algorithms for gathering autonomous robots
  to one point, where the robots locate in a simple
  triangle.
• Random algorithms for distributing autonomous
  robots to the boundary of a simple triangle such
  that the distance between any two neighbored
  robots is the same.
• Random algorithms for distributing autonomous
  robots to the rooms which locate one one floor
  such that the number of the robots in each room
  is the same.

15
DNA Computing
Limit of our silicon computer
It can not solve a large class of problems
(NP-hard problems) !
Example Hamiltonian Path Problem Given a graph
with n vertices and , find a path between two
given vertices such that the path involves every
vertex exactly once. (It costs about time.)
New computing paradigms

• Molecular computer DNA, RNA(ribonucleic
  acid),Protein
• Quantum computer
• Optical computer
• Bio computer cell, microbe

16
Why DNA Computing?

• Computation can be realized by designing
  operations on DNA molecules
• Large quantity of DAN stands can be provided in
  a short time with a small cost .
• DNA stands are much more stable than other
  living molecules

17
(No Transcript)
18
Adelmans experiment Solve Hamiltonian Path
Problem by DNA computing

• Algorithm
• Input Graph G with n vertices, among which are
  designated the input and output vertices u1 and
  u2.
• Generate paths in G randomly in large quantities.
  
• Reject all paths that do not begin with u1 and
  end in u2.
• Reject all paths that do not involve exactly n
  vertices.
• For each of the n vertices v, reject all paths
  that do not involve v.
• Output Yes if any path remains, No
  otherwise.

19
Generate paths

• Each vertex is associated with a random 20-mer
  strand DNA GGCTAGGTACCAGCATGCTT
  
• Each edge is associated with a 20-mer strand DNA
  which consists of the second and the first halves
  of the oligonucleotides encoding the vertices
  touching the edge.
  

vertex1 TATCGGATCGGTATATCCGA
vertex 2 GCTATTCGAGCTTAAAGCTA
edge(vertex1,vertex 2)GTATATCCGAGCTATTCGAG

• Generate all paths by annealing (Mix the
  oligonucleotides of vertices and edges)

20
DNA Computing Models
Algorithms on Adlemans model are designed by the
above three operations.
21
Our research and results

• Research topics
• DNA computational models.
• Paradigm of algorithm design for DNA computing.
  
• Logic and arithmetic operations with DNA
  strands.
• Results
• Procedures for logic and arithmetic operations
  with DNA strands.
• Algorithms for solving some graph problems with
  less DNA strands and lower error rate.
• Divide-and-conquer technique used for algorithm
  design in DNA computing.