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

Welcome to Scribd!

  • 66 views

    Solution Smith Rule

    Uploaded by

    Thu Ha Nguyen
    Single machine scheduling problems can be decomposed into simpler problems that can provide solutions for more complex models. This document discusses algorithms for two single machine scheduling problems: 1) minimizing total completion time (1||ΣCj), for which the optimal algorithm is shortest processing time (SPT) order; and 2) minimizing total weighted completion time (1||ΣwjCj), for which the optimal algorithm is weighted shortest processing time (WSPT) order. WSPT order sorts jobs in non-decreasing order of processing time divided by weight. An adjacent pairwise interchange proof shows WSPT order is optimal for 1||ΣwjCj.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    Solution Smith Rule

    Uploaded by

    Thu Ha Nguyen
    66 views3 pages
    Single machine scheduling problems can be decomposed into simpler problems that can provide solutions for more complex models. This document discusses algorithms for two single machine scheduling problems: 1) minimizing total completion time (1||ΣCj), for which the optimal algorithm is shortest processing time (SPT) order; and 2) minimizing total weighted completion time (1||ΣwjCj), for which the optimal algorithm is weighted shortest processing time (WSPT) order. WSPT order sorts jobs in non-decreasing order of processing time divided by weight. An adjacent pairwise interchange proof shows WSPT order is optimal for 1||ΣwjCj.

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Single machine scheduling problems can be decomposed into simpler problems that can provide solutions for more complex models. This document discusses algorithms for two single machine scheduling problems: 1) minimizing total completion time (1||ΣCj), for which the optimal algorithm is shortest processing time (SPT) order; and 2) minimizing total weighted completion time (1||ΣwjCj), for which the optimal algorithm is weighted shortest processing time (WSPT) order. WSPT order sorts jobs in non-decreasing order of processing time divided by weight. An adjacent pairwise interchange proof shows WSPT order is optimal for 1||ΣwjCj.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    66 views3 pages

    Solution Smith Rule

    Uploaded by

    Thu Ha Nguyen
    Single machine scheduling problems can be decomposed into simpler problems that can provide solutions for more complex models. This document discusses algorithms for two single machine scheduling problems: 1) minimizing total completion time (1||ΣCj), for which the optimal algorithm is shortest processing time (SPT) order; and 2) minimizing total weighted completion time (1||ΣwjCj), for which the optimal algorithm is weighted shortest processing time (WSPT) order. WSPT order sorts jobs in non-decreasing order of processing time divided by weight. An adjacent pairwise interchange proof shows WSPT order is optimal for 1||ΣwjCj.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 3

    Single machine problems

    Basic Scheduling Algorithms for Single Machine Problems


    Single machine model is the simplest type of scheduling models and a special case of all other environments. It is often found in practice when there is only one service point or a single stage manufacturing. Algorithms developed for the single machine model provide a basis for design of exact algorithms and heuristics for more complicated machine environments: multi-machine systems can often be decomposed into a number of single stage systems. In this lecture, we focus on the classical single machine problems with objective functions Cj, wjCj.

    1. Minimising total completion time: 1||Cj


    First let us develop an algorithm that finds an optimal schedule for 1||Cj. Solve the following instance with n=3 jobs. Job 1 2 3 pj 7 2 5

    To find its solution, let us generate all possible permutations of jobs and compute the completion times for the resulting schedules.

    10

    11 12 13 14 15 16 17

    18 19 20 21

    22 23

    Formulate the rule that solves the problem: _______________________________________________

    Single machine problems

    In the example above, the jobs in the optimal schedule are arranged in increasing (non-decreasing) order of their processing times. The rule that sorts the jobs in this order is known as SPT (shortest processing time), i.e., the next job to be scheduled will be the job with the smallest processing time. In general, take n jobs in order (1, 2, , n) and compute their completion times: C1 C2 Cn-1 Cn 1Cj = = p1 p1 p1 p1 np1 + p2 p2 + + p2 + + pn-1 pn-1 + Pn Pn

    1
    = = = + + +

    (n-1)p2 + + 2pn-1 +

    Thus, job 1 contributes np1, job 2 contributes (n-1)p2, and so on. If we want to minimize 1Cj, we want p1 to be the smallest, p2 the second smallest, etc. 25 7 7  7  7 j we need to sort the jobs in SPT order. As any 37
    7 974 7   48   9 8   6 8 58 9 9  7 9 9 7 other sorting, this requires O(nlog n) time.

    2. Minimising total weighted completion time: 1||1 wjCj


    Consider 1|| wjCj. Here - wj is the importance of job j, - wjCj characterises the total holding, or inventory costs incurred by the schedule. We start with the following instance of problem 1|pj=1| wjCj with n=3 jobs of unit length. Job 1 2 3 pj 1 1 1 wj 7 2 5

    Find an optimal sequence ____________________________________________ and the optimal value of the objective function ____________________________

    If all processing times are equal (pj=1) then the problem can be solved by ______________________ __________________________________________________________________________________

    Single machine problems

    For problem 1| |1wjCj intuitively we want to combine the rule above with SPT. It seems logical to sort the jobs in non-decreasing order of the ratios pj /wj. We call this rule WSPT (weighted SPT) rule or Smith's rule due to W.E. Smith who introduced it in 1956. We can prove Theorem 1. For 1|| wjCj, the WSPT rule is optimal.

    Proof (adjacent pairwise interchange argument) Suppose a schedule S, which is not WSPT, is optimal. In this schedule there must be at least two adjacent jobs i and k such that pi /wi > pk /wk. We assume that i precedes k and i starts at time t.
    k

    t+pi

    Swapping jobs i and k leads to a schedule S. We compare the value of the objective function F = wjCj for schedules S and S:
    F (S ) = F (S ' = )
    j i ,k

    w C
    j j j i ,k

    + wi C i + wk C k =
    ' + wi C i' + wk C k =

    w C

    F ( S ' F ( S ) = wi p k wk pi < 0. )

    This contradicts the optimality of schedule S and completes the proof of the theorem.

    t+pi +pk

    j i ,k

    w C
    j j j i , k

    + wi (t + pi ) + wk (t + pi + p k ) + wk (t + p k ) + wi (t + pi + p k )

    w C

    You might also like