greedy method report nhoel resos

Upload: nhoel-rs

Post on 13-Apr-2018

230 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    1/20

    Greedy Method

    Reported by: NHOEL C. RESOS

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    2/20

    The Greedy MethodTechnique

    The greedy methodis a general algorithmdesign paradigm, built on the following elements: confgurations: dierent choices, collections, or values

    to nd

    objective unction: a score assigned to congurations,which we want to either maximie or minimie

    !t wor"s best when applied to problems with thegreedy-choiceproperty:

    a globally#optimal solution can always be found by aseries of local improvements from a startingconguration$

    The Greedy Method 2

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    3/20

    %hat is a greedy algorithm&

    Greedy algorithm: 'an algorithm alwaysma"es the choice that loo"s best at themoment(

    )uman beings use greedy algorithms alot

    )ow to maximie your nal grade of thisclass&

    )ow to become a rich man&

    )ow does a casher minimie the number ofcoins to ma"e a change&

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    4/20

    %hat is a greedy algorithm&

    To guarantee that a greedy algorithm iscorrect,

    * things have to be proved:

    Greedy-choice property: 'we canassemble a globally optimal solution byma"ing locally greedy+optimal choices$(

    i$e$ The greedy choice is always part of certainoptimal solution

    Optimal substructure: 'an optimal solutionto the problem contains within it optimalsolutions to subproblems$(

    i$e$ global optimal solution is constructed fromlocal optimal solutions

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    5/20

    %hat is Greedy -lgorithm&

    !n the hard words: - greedy algorithmisan algorithm that follows theproblem solving heuristics of ma"ing the

    locally optimal choice at each stagewiththe hope of nding a global optimum$

    +src: http://en$wi"ipedia$org/wi"i/Greedy0algorithm

    1implify: 2hoose the best choice that3reachable4 at current state

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    6/20

    CHARACTERISTICS AND

    FEATURES To construct the solution in an optimal way. Algorithm

    aintains two sets!

    "#ne contains chosen items an$"The other contains re%ecte$ items.

    &ree$y algorithms ma'e good local choices in the hope that

    They result in!

    "An optimal solution.

    "Feasi(le solutions.

    6

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    7/20

    C#NTINUED)

    The gree$y algorithm consists o* *our +,- *unction.

    A *unction that chec's whether chosen set o* items proi$e a

    solution. A *unction that chec's the *easi(ility o* a set.

    The selection *unction tells which o* the items is the most

    promising.

    An o(%ectie *unction! which $oes not appear e/plicitly! giesthe alue o* a solution.

    7

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    8/20

    #0TII1ATI#N 0R#23ES

    An optimi4ation pro(lem5

    &ien a pro(lem instance! a set o* constraintsan$ anobjective function.

    Fin$ a feasiblesolution *or the gien instance *or which theo(%ectie *unction has an optimal alue.

    Either ma/imum or minimum $epen$ing on the pro(lem (eingsole$.A *easi(le solution that $oes this is calle$ optimalsolution.

    8

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    9/20

    Continue$)

    Feasible5A *easi(le solution satis*ies the pro(lem6s constraints

    Constraints5The constraintsspeci*y the limitations on the re7uire$solutions.

    9

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    10/20

    1ample 5sage of Greedy

    6or better explanation we use oldsimple problem: Travelling 1alesman7roblem:

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    11/20

    T17

    The 7roblem is how to travel from city -and visit all city on the map, then bac" tocity - again$

    The rules: you only visit each city once andyou can4t pass through any traversed path$

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    12/20

    1olution:

    6ind the shortest path from city -+start toany other city$

    8ecause the nearest city is 8, so we go to 8

    8-

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    13/20

    6rom 8, we nd any other city but

    -+because - has been visited that hasnearest path$ 1o we choose 2:

    9eep tuning on

    8-

    2

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    14/20

    6rom 2, we loo" to nearest city again,

    but don4t loo" for - and 8, becauseboth has been visited$ 1o we choose ;$

    1oon end

    8-

    2 ;

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    15/20

    -t this node+;, we can4t go to any city,because all neighbor of ; has beenvisited$ %e go bac" to rst city+-$

    -nd that was how to solve T17 problem$

    8-

    2 ;

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    16/20

    1ample 5sage of Greedy

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    17/20

    The knapsack problemor rucksack problemis a problem in combinatorial optimiation: Given a set of items,each with a weight and a value, determine the number of each item to include in a collection so that the total

    weight is less than or equal to a given limit and the total value is as large as possible$ !t derives its name from theproblem faced by someone who is constrained by a xed#sie "napsac"and must ll it with the most valuableitems$The problem often arises in resource allocationwhere there are nancial constraints and is studied in elds suchas combinatorics, computer science, complexity theory, cryptography, applied mathematics, anddaily fantasy sports$The "napsac" problem has been studied for more than a century, with early wor"s dating as far bac" as ?$@

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    18/20

    -dvantage of Greedy

    Greedy is easy to be implemented$ust search the best choice from thecurrent state that 3reachable4 +has

    any paths or any connections$

    !n simple case, greedy often give youthe best solution$

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    19/20

    ;rawbac" of Greedy

    !n large and complex case, greedydoesn4t always give you the bestsolution, because it4s Hust search and

    ta"e the best choice that you canreach from the current state$

    !t ta"es longer time than any other

    algorithms for big case of problem

  • 7/26/2019 Greedy Method Report NHOEL RESOS

    20/20

    0R#S AND C#NS

    0R#S5 They are easier to implement! they re7uire much less computing resources! they are much *aster to e/ecute.

    &ree$y algorithms are use$ to sole optimi4ation pro(lems

    C#NS5 Their only $isa$antage (eing that they not always reach the

    glo(al optimum solution8 on the other han$! een when the glo(al optimum solution isnot reache$! most o* the times the reache$ su("optimalsolution is a ery goo$ solution.

    2-