Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 0 views

    Partiel-RO-2015-21-04

    Uploaded by

    Georgio Karam
    cvvvvvcvvvv

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Partiel-RO-2015-21-04

    Uploaded by

    Georgio Karam
    0 views2 pages
    cvvvvvcvvvv

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    cvvvvvcvvvv

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    0 views2 pages

    Partiel-RO-2015-21-04

    Uploaded by

    Georgio Karam
    cvvvvvcvvvv

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    Département : Génie Electronique

    UNIVERSITE LIBANAISE Semestre : 8ème


    FACULTE DE GENIE Date : 21 Avril 2015
    BRANCHE I Professeur : Dr. Nada Chendeb

    Documents interdits Durée : Deux heures

    Examen Partiel
    Recherche Operationnelle (2.8.5.T - 2.8.5.C)
    Exercise 1:
    We want to assign 5 tasks to 5 machines. Assignment costs are given in the following
    table:

    Machine 1 Machine 2 Machine 3 Machine 4 Machine 5


    Task 1 15 40 5 20 20
    Task 2 22 33 9 16 20
    Task 3 40 6 28 0 26
    Task 4 8 0 7 25 60
    Task 5 10 10 60 15 5

    Apply the appropriate algorithm to find the assignment that leads to the minimum
    cost. What is this cost? Find alternative solutions if any.

    Exercise 2:
    Consider the queue of a supermarket till. The number of persons waiting in the queue
    constitutes a random set of variables (Xn) n € N where n designates the number of
    minutes passed from the opening of the till. We suppose that Xn evolves as follows:
    – If the queue is not empty, a new client passes to the till with a probability p.
    – If the queue contains 5 persons, people don't like this and no one comes to
    wait in this queue.
    – If the queue contains strictly less than 5 persons, a new person comes into the
    queue with probability q.
    1. Show that the set (Xn) n € N is a discrete Markov chain with 6 states.
    2. Demonstrate that:
    P(Xn+1=k+ 1 | Xn=k) = q(1−p), 0≤k≤4
    P(Xn+1=k−1 | Xn=k) = p(1−q), 1≤k≤5
    3. Calculate all other non-null transition probabilities and write the complete
    transition matrix Q.
    4. Show that the Markov chain is irreducible. Is it a birth death process?
    5. Calculate the stationary probabilities for p = q = 1/2.
    6. What is the average return time to empty queue? Use this
    formula E (T i / X 0  i )  1 /  (i )

    Exercise 3
    A company needs the same electronic data processing capacity at n = 10 different
    locations:
    • The number of computational jobs within a fixed time is ε, and is poisson distributed
    with the parameter λ = 4 jobs/s.
    • The number of computational operations per job is assumed to be exponentially
    distributed with a mean value α = 5.4*105 operations/job.

    Three different computer installations shall be compared:


    A. One workstation at each location that perform LA = 3·106 op/s.
    B. At a central location, a workstation pool is installed. Each workstation is
    available for jobs from any location if it is free. The performance of one
    workstation is the same as in A.
    C. At a central location, a faster computer with LC = 2.4*107 op/s is installed.
    This computer executes the jobs one after another.

    All waiting jobs shall be executed with FIFO strategy. At the beginning, assume that
    there will be no losses in the queues.

    a) What kind of queuing systems do we have in A, B and C? Find out the


    parameters of the stochastical input/output processes and the traffic load ρ.
    How many workstations are needed in version B?
    b) Develop analytically the formulas of the response time of the queuing systems
    in A, B and C.
    c) What is the mean processing time (response time) of a job in version A, B and
    C?
    d) Assume that the queues of all three versions have N = 30 input buffer places.
    What is the probability of a loss? Develop your equations.
    e) Analyze and compare the three systems

    Note: If you are unable to develop your equations, you can use the sheet attached
    to this exam, part of the grade will be lost.

     Good luck 

    You might also like