OM Week 9

Project

Series of related activities directed toward a major output and requiring a significant period of time to perform

Project Management

Defining, planning, directing and controlling resources so the project meets its technical, cost, and time constraints

Life Cycle of a Project

1. Define

• Goal and scope

1. Plan

• Activities and timing

1. Execute

• Do the work

1. Close

• Deliver and learn

What are the two planning tools for projects?

1. Gantt Chart

2. Network Planning Diagram

What are the strengths of a Gantt Chart?

• Great for calendar visibility

• Shows starts, finishes, overlaps

• best for communicating the schedule

What is a Gantt Chart

A bar chart of tasks over time

What is a Network Planning Diagram?

A flow-chart-like map of activities

What are the strengths of a Network Planning Diagram?

• Great for logic visibility

• Shows precedence constraints directly

• Best for critical path analysis

What planning tool should we use when we ask the question, “When does each task happen?”

Gantt chart

What planning tool should we use when we ask the question, “What must happen before what?”

Network Planning Diagram

What is the main limitation of a Gantt Chart?

We can see when but not why

On a Network Planning Diagram what do the following represent:

• Nodes

• Arrows

• The longest path

• Nodes = activities

• Arrows = “must finish before”

• The longest path = project duration

Path

One complete route through the network from the first activity (A) to the last activity (K), following the arrows

Critical Path

The longest path

Which path is the bottleneck

The longest path

Why is the critical path the project duration?

The project can’t launch until every path is done

Why does Critical Path Method (CPM) matter for a Project Manager (4 reasons)?

1. Not all tasks deserve equal attention: CPM quantifies exactly how many days each task can slip before the project is in trouble (That’s where you focus)

2. Resource allocation: CPM tells you which activities are critical (zero slack) and which activities have buffers

3. Spending decisions: When you need to speed up, CPM tells you where to spend money (paying to rush a task that already has slack is wasted budget)

4. Scheduling flexibility: For every task, CPM gives you an earliest and latest window (You can shift non-critical tasks to balance workload or manage cash flow without delaying the festival)

Path enumeration tells you the critical path, CPM tells you

how critical every ask is,

What are the three steps in the CPM Game Plan?

1. Forward pass: work left to right to find the earliest start and finish for every activity

2. Backward pass: work right to left to find the latest start and finish for every activity

3. Compute slack: the difference tells you how much breathing room each task has

• Zero slack = critical

What do the four numbers of each activity mean?

Forward-pass Labels (“earliest”)

• ES (Earliest Start): soonest this task can begin

• EF (Earliest Finish): EF + ES + duration

Backward-pass Labels (“latest”)

• LS (Latest Start): Latest you can begin without delaying the festival

• LF (Latest Finish): Latest this task can end → LS = LF - Duration

Forward pass

“Who am I waiting for?”

• Wait for the slowest → take the max

Backward pass

“Who am I making wait?”

• Meet the tightest deadline → take the min

What question does Forward Pass answer (SXSW context)?

“If everyone starts as early as possible, what is the earliest each SXSW task cans tart and finish?”

What are the two rules of Forward Pass?

1. ES = max(EF of all predecessors)

• A task can’t start until every predecessor is done.

• If a task has multiple predecessors, take the maximum - you’re waiting for the slowest one

2. EF = ES + duration

• Once you know when a task starts, add its duration to get when it finishes

What do you do when there are multiple predecessors?

Wait for the slowest one and take the max

Max Rule

You can’t start “cooking” until ALL ingredients arrive - the last delivery sets the start time

At every node ES =

Max of predecessor EFs

• EF = ES + duration

The earliest the project can launch is equal to

The earliest finish of the final activity

What question does Backward Pass answer (SXSW context)?

“The festival must launch on day 46. What is the latest each task can start and finish without pushing that date?”

What are the two rules of Backward Pass?

1. LF = min(LS of all successors)

• A task must finish in time for every task that depends on it

• If multiple tasks follow, take the minimum (you need to be done before the earliest successor starts)

2. LS = LF - duration

• Once you know when a task must finish, subtract its duration to get when it must start

Min Rule

You promised two people deliverables. The one who needs it sooner sets your deadline

If you start any task after its LS, the festival launches [BLANK]

• Late/Early

Late

Slack

How many days you can delay an activity without delaying the entire project

Slack Formula

Slack = LS - ES = LF - EF

• Both give the same answer (Pick whichever is easier)

Slack

How many days you can delay an activity without delaying the entire festival

Why is an activity critical if there is zero slack

There is no room for delay

• Any slippage pushes back the projects launch

Project Crashing

Deliberately spending extra money to shorten specific activities

• Hire overtime crews, pay rush fees, use express shipping

Why do only critical activities matter when needing to speed up a certain portion of a project?

Shortening a noncritical task just adds to its slack → the project will not launch sooner

Each activity has a minimum duration, therefore

You can only speed it up so much

Each day of speedup has a

Crash Cost

• A technique to shorten a project's duration by adding resources to critical path activities, minimizing extra cost

After you crash one activity

The critical path can change

Crash the cheapest options first AND

Re-check after every change

Crash Cost per Day

The price to speed up that activity by one day

Some activities can’t be crashed at all (True/False)

True

Step-by-step Method for Crashing

1. Identify the current critical path(s)

2. List the crashable critical activities and their cost per day

3. If there is one critical path: crash the cheapest critical activity by one day

4. If there are multiple critical paths: crash an activity shared by all critical paths (If no single activity works, you may need to crash one activity on each path simultaneously)

5. Recompute all path lengths, check whether a new path has become critical

6. Repeat until you hit the target duration or run out of crashable activities

Why do we recompute after each crash?

The critical path can change and ignoring that wastes money

Why is Project crashing an economic decision and not just a scheduling decision?

Every day you shave off has a price tag → the algorithm ensures you spend as little as possible