Modeling%20Tough%20Scheduling%20Problems%20with%20Software

1
Modeling Tough Scheduling Problems with Software

• Alex S. Brown
• http//www.alexsbrown.com
• Mitsui Sumitomo Insurance Group, USA

2
Objectives

• Get past frustration (intermediate)
• Learn new tools and techniques (expert)
• Not for beginners
• Software-brand neutral
• Interactive, case-study format

Participate!
3
Key Terms

• Network Diagram
• Gantt Chart
• Two Types of Projects

Serial Dependencies
Parallel Opportunities
4
Scenario

• Six Tasks
• Three Resources

5
Five Directions - Pick One!

• Resource Leveling
• Representing Hard Soft Dependencies
• Only Full-Time Work
• Quickly-Changing Work Assignments
• Maintaining Schedules During Execution

Go!
Go!
Go!
Go!
Go!
6
Resource Leveling - Key Data

• Gather complete data at the start
• Resource start dates and availability
• Vacations
• Accurate work and duration estimates
• Accurate sequence relationships
• Good data is essential
• When data is good enough, level to discover the
  real end date

7
Principles of Leveling

• Find a repeatable method
• Level regularly
• Changes in any estimate or assumption may require
  re-leveling
• Leveling is iterative
• Methods should allow small, frequent adjustments

8
Scheduling Tools for Leveling

• View resource use over time
• Adjust key task info
• Start dates
• Delays on tasks
• Allocation of work over time by task
• Resource availability vs. Work assigned

9
Breakout Discussion

• Level the scenario for 40-hour weeks, 8-hour days
  for each resource
• Different approaches on the handouts
• Key questions
• What method do you normally use?
• Which is easiest to set up?
• Which is easiest to maintain?
• Five minutes for discussion

10
Possible Solutions

• Stretch all work to five days duration
• Make work sequential
• Move start dates
• Add delays
• Break down the task list differently to simplify
  leveling

11
Selecting a Technique

• What is easiest in your software?
• How hard is maintenance?
• What methods allow for automation for all or part
  of the work?
• What is the impact on reporting and analysis
  techniques?
• Critical path
• Earned value

12
Questions Answers on Leveling
Pick Next Topic
13
Hard Soft Dependencies

• Hard dependencies MUST be obeyed
• Soft dependencies SHOULD be obeyed
• Real-life schedules often break or change soft
  dependencies to achieve required end dates
• Challenge your assumptions is it really
  IMPOSSIBLE for task B to start until task A ends?

14
Issues with Dependencies

• Software usually recognizes only hard
  dependencies
• Critical path analysis requires an accurate
  network diagram
• Documenting dependencies is
• Challenging
• Absolutely essential

15
Breakout Discussion

• Scenario uses soft dependencies
• Soft dependencies driven by resource limitations
• Questions
• Where do you document dependencies?
• What is the critical path?
• What do you want the critical path to be?
• How does the schedule respond to change?

16
Documenting Dependencies

• Many options
• Two network diagrams - hard only and current
• Narrative explanation of dependencies
• Separate columns or spreadsheets track hard and
  soft dependencies
• Pick a method you can maintain easily

17
Dependencies in Software

• Scheduling software usually enforces dependencies
  strictly (all hard)
• Options
• Only enter hard dependencies - satisfy soft
  dependencies during leveling
• Enter both hard and soft dependencies for each
  task, update as needed
• Enter only the constraining dependency

18
Consequences of Approaches

• Calculation of critical path
• Clarity of network diagram
• Only hard is clearest diagram
• Hard soft creates overlapping lines
• Only constraining looks like a serial projects
  network
• Respond to change differently

19
Questions Answers on Dependencies
Pick Next Topic
20
Only Full-Time Work

• Reasons to insist on full-time work assignments
• Use critical chain method
• Improve control of work authorization
• Clarify status reports
• New problems
• Adds more tasks
• Can make the schedule more complex

21
Breakout Discussion

• Initial scenario changed
• Resource A and C have clearer assignments
• Answer these questions
• Why does the work take longer now?
• What resource-leveling methods do you prefer for
  these problems?

22
What does 50 allocated mean, anyway?

• Example illustrates a possible trap
• Two, five-day tasks cannot be done in five days
  by one person
• (most of the time)
• Always need to add common sense to any schedule,
  and not rely solely on software
• True with ANY methodology

23
Questions Answers on Full-Time Work
Pick Next Topic
24
Quickly-Changing Schedules

• Real schedules change every reporting period
• Estimate changes - work duration
• Assignment changes
• Resources added, removed, and rescheduled
• Resource availability changes
• Optimize schedule, make new forecasts
• The only certainty is change itself

25
Planning for Changes in the Plan

• Predict future changes
• What is likely to change?
• What changed often in the past?
• What project risks will cause change?
• Create a small schedule with typical tasks
• Try making typical changes
• Experiment to find best techniques

26
Breakout Discussion

• Start with initial scenario
• Apply changes
• Resources no longer available
• Resources available more hours per week
• Estimate changes
• Build a new model and repeat
• Hard and soft dependencies
• Different leveling techniques

27
Change is Fundamental

• Experiment with change in a small schedule
• Best way to evaluate and compare different
  techniques
• Great way to explore, understand, and evaluate
  scheduling software
• Best techniques will vary -- find the best for
  your schedules, for your software

28
Change Drives Automation

• Change is difficult with hand-drawn schedules and
  budgets
• Software updates every view of a schedule
  simultaneously
• WARNING if the model is poor, software makes
  changes HARDER
• Understand basic principles BEFORE automating

29
Questions Answers on Quickly-Changing Schedules
Pick Next Topic
30
Maintenance During Execution

• Key steps
• Apply actual work
• Update estimates for remaining work
• Optimize the schedule and update forecasts
• Many managers shelve schedules after developing
  them
• Schedule maintenance key to learning as an
  organization and as a manager

31
Benefits of Maintenance

• Compare actuals to estimates
• Improve estimates for next time
• Improve control of work and scope
• Required for earned value and benchmarking
• Accurate, up-to-date forecasts of delivery dates
• Prove scheduling techniques

32
Package-Specific Features

• Timesheets automate and distribute entry of
  actuals
• Entry of actuals vary by software package
• Different packages reschedule remaining work
  differently
• Do not focus on software differences, focus on
  creating a new forecast

33
Breakout Discussion

• Review actual progress after week 1
• Schedule the remaining work
• Questions
• Can you predict how your software would
  reschedule the remaining work?
• Do we change dependencies now that work is
  started?

34
What SHOULD Software Do?

• Actuals sometimes break rules in the schedule
• Actual work starts before predecessor is complete
• Must meet deadlines missed
• Software packages make assumptions and raise
  warnings each has a different approach

35
Responsibility As Software Users

• Understand that actual work might break our model
• Software may react unpredictably when our models
  are wrong
• Understand software settings, defaults, and
  options
• Experiment, experiment, experiment create small
  schedules to test and learn

36
Important Choices

• How you build a schedule determines how you
  maintain it
• Schedule time to update the schedule
• Forecast new end dates regularly
• Larger projects mean
• Maintenance is critical
• Maintenance is difficult

37
Questions Answers on Schedule Maintenance
Pick Next Topic
38
Review - Five Tough Problems

• Resource Leveling
• Representing Hard Soft Dependencies
• Only Full-Time Work
• Quickly-Changing Work Assignments
• Maintaining Schedules During Execution

39
Resources Available On-Line

• Additional exercises from 2003 session
• On-line paper
• Comparison of novice vs. expert use of scheduling
  software
• Ideas to improve your skills
• References and additional reading

40
An Appeal to Scheduling Teachers

• More advanced classes
• Teach techniques from real examples
• Spend MORE time on resource-leveling than on
  dependency types
• Incorporate tough problems into your courses,
  even introductory ones
• Encourage students to experiment often with
  small, test schedules

41
Complete Your Learning Experience

• This presentation is a first step
• Run exercises using your software
• Select scenarios from your schedules and try new
  techniques
• Hear it and you will forget it
• Do it and you will remember it!
• Share with your coworkers

42
Change Your Relationship with Your Scheduling
Software

• No need to get angry any more

43
(No Transcript)
44
Questions Answers

• Contact speaker at
• http//www.alexsbrown.com
• alexsbrown_at_alexsbrown.com
• alexbrown_at_msigusa.com
• Share ideas, examples, exercises!

45
Modeling Tough Scheduling Problems with Software

• Alex S. Brown
• http//www.alexsbrown.com
• Mitsui Sumitomo Insurance Group, USA