Lecture7

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: . All tasks with zero total float are critical. 2 . All critical tasks have zero free float. 3 . All critical tasks have zero total float. 4 Statements 2-4 are all true. Statement 1 is not true by the following example: C has zero free float but 1 total float. C (8) S=[5,6] F=[13,14] START A (5) S=[0,0] F=[5,5] F (11) S=[13,14] F=[24,25] D (8) S=[5,5] F=[13,13] G (1) S=[13,24] F=[14,25] END H (12) S=[13,13] F=[25,25] J. G. Carlsson, U of MN ISyE Lecture 7: PERT & CPM October 3/17, 2013 27 / 38 Crashing Crashing an activity refers to taking special costly measures to reduce the duration of an activity below its normal value E.g. “if we pay \$50,000, we can reduce the duration of activity D from 8 to 7” Naturally, it is only worth considering crashing activities on the critical path Crashing may change the critical path by reducing the duration of critical activities J. G. Carlsson, U of MN ISyE Lecture 7: PERT & CPM October 3/17, 2013 28 / 38 Crashing as an LP Suppose that each activity i can be crashed by paying pi to reduce the duration by 1 day (or whatever time unit we like) If we want to crash by xi days, then we pay pi xi We have a given deadline that we have to meet: how can we crash in the cheapest way possible? B (3) pB = 3 C (8) pC = 8 START D (11) pD = 8 E (1) pE = 4 A (5) pA = 1 0 END F (12) pF = 6 J. G. Carlsson, U of MN ISyE Lecture 7: PERT & CPM October 3/17, 2013 29 / 38 Crashing as an LP Say we need to complete by 15 days: minimize 10xA + 3xB + 8xC + 8xD + 4xE + 6xF yA yB s.t. ≥ 0 + 5 − xA ≥ yA + 3 − xB yC yD ≥ yA + 8 − xC ≥ yB + 11 − xD yD yE yF ≥ yC + 11 − xD ≥ yC + 1 − xE ≥ yC + 12 − xF yA , yB , yC , yD , yE , yF xA , xB , xC , xD , xE , xF ≤ 15 ≥0 yA , yB , yC , yD , yE , yF ≥0 Here the yi ’s represent completion times and the xi ’s represent how much we crash J. G. Carlsson, U of MN ISyE Lecture 7: PERT & CPM October 3/17, 2013 30 / 38 Optimal solution At right, a “cost-duration tradeoff curve” J. G. Carlsson, U of MN ISyE Lecture 7: PERT & CPM October 3/17, 2013 31 / 38 PERT Networks It is natural to try to take some uncertainty into account when designing project networks A PERT network estimates times based on three input values: The optimistic time a occurs if execution goes extremely well The most likely time m is just what it sounds like The pessimistic time b occurs if execution goes extremely poorly ¯ The average duration time D and the variance ν are defined by ¯ D ν J. G. Carlsson, U of MN ISyE a + 4m + b 6 ( )2 b−a = 6 = Lecture 7: PERT & CPM October 3/17, 2013 32 / 38 PDM The precedence diagramming method (PDM) is an extension of PERT/CPM in which mutually dependent activities can be performed partially instead of sequentially In PERT/CPM, we assume that an activity must be completely finished before we can work on its successors PDM generalizes this by allowing the following additional relationships: SSAB (start-to-start) lead: activity B cannot start until activity A has been in progress for at least SS time...
View Full Document

This document was uploaded on 02/23/2014 for the course MANAGMENT 2201 at University of Michigan.

Ask a homework question - tutors are online