π Lecture 44 β PROJECT MANAGEMENT (Contd.)
π Overview: This lecture continues the study of project management, moving from network diagrams and project life cycles into the critical areas of time estimation and variance analysis. It covers both deterministic and probabilistic time estimates, the computational algorithms for Early Start, Early Finish, Late Start, and Late Finish, and introduces project crashing, time-cost trade-offs, risk management, and operations strategy.
ποΈ Topics Covered
The lecture covers two types of time estimates (deterministic and probabilistic), the computing algorithm for forward and backward path analysis using ES, EF, LS, and LF, probabilistic time estimates including optimistic, pessimistic, and most likely times along with expected time and variance calculations, path probabilities using Z-scores, time-cost trade-offs through project crashing, project management software tools, project risk management, and the development of a project management-based operations strategy.
π Lecture Summary
Time Estimates
There are two common types of time estimates in project management. Deterministic time estimates are used when activity times are fairly certain. Probabilistic time estimates allow for variation in activity durations, using three different time estimates to account for uncertainty.
The hospital example is used to illustrate how activities from locating the facility to making the hospital operational are represented in a network diagram. Students are encouraged to practice drawing network diagrams using both activity-on-node and activity-on-arrow conventions.
Computing Algorithm
Network activities are analyzed using four key time parameters: ES (Early Start), EF (Early Finish), LS (Late Start), and LF (Late Finish). These are used to determine the expected project duration, slack time, and the critical path.
π Definition β Forward Pass: The process of calculating ES and EF by moving forward through the network from start to finish. π Definition β Backward Pass: The process of calculating LS and LF by moving backward through the network from finish to start. π Definition β Slack Time: The amount of time an activity can be delayed without delaying the entire project, calculated as LS - ES or LF - EF. π Definition β Critical Path: The longest path through the network diagram, which determines the minimum project completion time; activities on this path have zero slack.
π Formula β EF = ES + Activity Time π Formula β LS = LF - Activity Time
Using the hospital network diagram with activities and durations (Locate facilities: 8 weeks, Order machines: 6 weeks, Machine setup: 3 weeks, Interview: 4 weeks, Hire and train medical staff: 9 weeks, Remodel: 11 weeks, Operational: 1 week), the forward and backward passes are performed to identify the critical path and slack times.
Probabilistic Time Estimates
Probabilistic estimates use three time estimates for each activity. Optimistic time (tβ) is the time required under optimal conditions. Pessimistic time (tβ) is the time required under worst conditions. Most likely time (tβ) is the most probable length of time that will be required.
From these three estimates, two important parameters are calculated: the Expected Time (tβ) and the Variance (ΟΒ²).
π Formula β Expected Time: tβ = (tβ + 4tβ + tβ) / 6 π Formula β Variance: ΟΒ² = (tβ - tβ)Β² / 36
π Definition β Variance: The square of the standard deviation of activities on a path. The size of the variance reflects the degree of uncertainty associated with an activity's time; the larger the variance, the larger the uncertainty.
π‘ Why this matters: Probabilistic time estimates allow project managers to account for uncertainty in activity durations, providing a more realistic view of project completion times and enabling better risk assessment.
Path Probabilities
The Z-value indicates how many standard deviations of the path distribution a specified time is beyond the expected path duration.
π Formula β Z = (Specified time - Path mean) / Path standard deviation
If the value of Z is +2.50 or more, the probability of path completion by the specified time can be treated as 100 percent.
Time-cost Trade-offs: Crashing
Crashing is the shortening of activity duration by paying more money to complete a project more quickly. Since the critical path determines the length of a project, it makes sense to reduce the length of activities on the critical path.
Procedure for crashing:
- Crash the project one period at a time
- Only crash an activity on the critical path
- Crash the least expensive activity first
- If there are multiple critical paths, find the sum of crashing the least expensive activity on each critical path
Critical path activities should be reduced until the project is reduced to the desired length or until the cost per day of crashing exceeds the savings per day. If there are multiple critical paths, they must be shortened simultaneously.
π Example: The graph of Total Cost vs. Project Duration shows that as the project is shortened (crashed), direct costs increase while indirect costs decrease. The optimal project duration occurs at the point where the total cost (cumulative cost of crashing plus expected indirect costs) is minimized.
Example: Reforestation Pilot Project
A manager is undertaking a reforestation pilot project with six activities:
| Activity | Precedes | Optimistic (a) | Most Likely (m) | Pessimistic (b) |
|---|---|---|---|---|
| Start | U, V | - | - | - |
| U | W | 35 | 50 | 65 |
| V | W, X | 28 | 40 | 52 |
| W | Z | 26 | 35 | 44 |
| X | Y | 28 | 40 | 52 |
| Y | Z | 26 | 29 | 38 |
| Z | End | 36 | 60 | 84 |
Step 1: Calculate Expected Time (t) for each activity
- Activity U: (35 + 4(50) + 65)/6 = 300/6 = 50 days
- Activity V: (28 + 4(40) + 52)/6 = 240/6 = 40 days
- Activity W: (26 + 4(35) + 44)/6 = 210/6 = 35 days
- Activity X: (28 + 4(40) + 52)/6 = 240/6 = 40 days
- Activity Y: (26 + 4(29) + 38)/6 = 180/6 = 30 days
- Activity Z: (36 + 4(60) + 84)/6 = 360/6 = 60 days
Step 2: Calculate Standard Deviation (Ο) for each activity
- Activity U: (65-35)/6 = 5 days
- Activity V: (52-28)/6 = 4 days
- Activity W: (44-26)/6 = 3 days
- Activity X: (52-28)/6 = 4 days
- Activity Y: (38-26)/6 = 2 days
- Activity Z: (84-36)/6 = 8 days
Step 3: Identify the Critical Path Three possible paths from Start to End:
- Start-U-W-Z-End = 50 + 35 + 60 = 145 days
- Start-V-X-Y-Z-End = 40 + 40 + 30 + 60 = 170 days (Critical Path)
- Start-V-W-Z-End = 40 + 35 + 60 = 135 days
The critical path is Start-V-X-Y-Z-End with an expected project duration of 170 days.
Step 4: Calculate Portfolio Standard Deviation ΟΒ² for the critical path = (4)Β² + (4)Β² + (2)Β² + (8)Β² = 16 + 16 + 4 + 64 = 100 Ο for the project portfolio = β100 = 10 days
(Note: The portfolio Ο of 10 days is less than the sum of individual standard deviations of 18 days, verifying the calculation is correct.)
Step 5: Calculate Probability of Completion within 200 Days π Formula β Z = (X - ΞΌ) / Ο = (200 - 170) / 10 = 30/10 = 3.0
According to the standard normal table, the area at Z = 3.0 is 0.4987. Adding 0.5 for the left-hand side of the curve gives 0.9987.
Results from Example Questions:
- Expected time for Project Completion: 170 days
- Slack time for activity W: 25 days
- Probability of completing critical path within 200 days: 0.9987
- Expected time for activity Y: 30 days
- Standard deviation for activity Z: 8 days
Solved Example: Kabbadi Stadium Project
A construction project with the following activity time estimates (in days):
| Activity | Optimistic | Most Likely | Pessimistic | Immediate Predecessor |
|---|---|---|---|---|
| A | 1 | 4 | 7 | - |
| B | 2 | 6 | 7 | - |
| C | 3 | 3 | 6 | B |
| D | 6 | 13 | 14 | A |
| E | 3 | 6 | 12 | A, C |
| F | 6 | 8 | 16 | B |
| G | 1 | 5 | 6 | E, F |
Step 1: Calculate Expected Time and Variance
| Activity | Expected Time | Variance |
|---|---|---|
| A | 4.00 | 1.00 |
| B | 5.50 | 0.69 |
| C | 3.50 | 0.25 |
| D | 12.00 | 1.78 |
| E | 6.50 | 2.25 |
| F | 9.00 | 2.78 |
| G | 4.50 | 0.69 |
Step 2: Forward Pass (ES and EF)
| Activity | ES | EF | t |
|---|---|---|---|
| A | 0.00 | 4.00 | 4.00 |
| B | 0.00 | 5.50 | 5.50 |
| C | 5.50 | 9.00 | 3.50 |
| D | 4.00 | 16.00 | 12.00 |
| E | 9.00 | 15.50 | 6.50 |
| F | 5.50 | 14.50 | 9.00 |
| G | 15.50 | 20.00 | 4.50 |
Step 3: Backward Pass (LS and LF)
| Activity | LS | LF | t |
|---|---|---|---|
| G | 15.50 | 20.00 | 4.50 |
| F | 6.50 | 15.50 | 9.00 |
| E | 9.00 | 15.50 | 6.50 |
| D | 8.00 | 20.00 | 12.00 |
| C | 5.50 | 9.00 | 3.50 |
| B | 0.00 | 5.50 | 5.50 |
| A | 4.00 | 8.00 | 4.00 |
Step 4: Determine Slack and Critical Path
| Activity | ES | LS | EF | LF | Slack | Critical? |
|---|---|---|---|---|---|---|
| A | 0.00 | 4.00 | 4.00 | 8.00 | 4.00 | No |
| B | 0.00 | 0.00 | 5.50 | 5.50 | 0.00 | Yes |
| C | 5.50 | 5.50 | 9.00 | 9.00 | 0.00 | Yes |
| D | 4.00 | 8.00 | 16.00 | 20.00 | 4.00 | No |
| E | 9.00 | 9.00 | 15.50 | 15.50 | 0.00 | Yes |
| F | 5.50 | 13.00 | 14.50 | 15.50 | 1.00 | No |
| G | 15.50 | 15.50 | 20.00 | 20.00 | 0.00 | Yes |
The critical path is B-C-E-G with a total expected time of 20 days.
Step 5: Calculate Probability of Completion within 23 Days Path variance for B-C-E-G = 0.69 + 0.25 + 2.25 + 0.69 = 3.89 Ο = β3.89 = 1.972 Z = (23 - 20) / 1.972 = 3 / 1.972 = 1.5210
Using the Normal Distribution table, the probability of completing the project in 23 days is 0.9357.
Project Management Software Tools
Several software tools are used in project management:
- Computer Aided Design (CAD)
- Groupware (e.g., Lotus Notes)
- Project Management Software: CA Super Project, Harvard Total Manager, MS Project, Sure Track Project Manager, Time Line
Advantages of PM Software:
- Imposes a methodology
- Provides logical planning structure
- Enhances team communication
- Flags constraint violations
- Provides automatic report formats
- Offers multiple levels of reports
- Enables what-if scenarios
- Generates various chart types
Project Risk Management
π Definition β Risk: The occurrence of events that have undesirable consequences, such as delays, increased costs, inability to meet specifications, or project termination.
Risk Management involves four steps:
- Identify potential risks
- Analyze and assess risks
- Work to minimize occurrence of risk
- Establish contingency plans
Operations Strategy
Organizations often set up a separate Project Management department or cell to administer unique and non-repetitive activities. The scope of the project determines whether to use project management software. Project teams typically operate as matrix teams, with employees from different functional departments working with the project team. In such situations, organizations devise a strategy where the project manager should lead the team, as they are more aware of the situation faced by the whole organization and the constituent functional departments.
β Key Takeaways
The critical path is the longest path through the network and determines the minimum project completion time; activities on this path have zero slack and must be closely monitored. Probabilistic time estimates using optimistic, most likely, and pessimistic times allow for the calculation of expected time and variance, enabling probability analysis of project completion within a specified time using Z-scores. Project crashing involves shortening the project by reducing activities on the critical path, starting with the least expensive option, and if multiple critical paths exist, they must be shortened simultaneously. The standard deviation of the project portfolio is calculated as the square root of the sum of variances along the critical path, not the sum of individual standard deviations. Risk management and appropriate software tools are essential for successful project execution, and organizations typically use matrix structures where the project manager leads cross-functional teams.
π§ Quick Revision Questions
- What are the two types of time estimates in project management and when is each used?
- How do you calculate the expected time (tβ) and variance (ΟΒ²) for an activity using probabilistic time estimates?
- What is the procedure for crashing a project, and why is it important to crash the least expensive activity on the critical path first?
- If a project has a critical path with expected duration of 170 days and standard deviation of 10 days, what is the probability of completing it within 180 days?
- What are the four steps of risk management in project management?
π Lecture 45 β WAITING LINES
π Overview: This lecture explores the formation and management of waiting lines (queues) in service systems, even when those systems are underloaded at the macro level. It introduces queuing theory as a mathematical approach to balance service capacity costs with customer waiting costs, covering system characteristics, performance measures, and both mathematical and non-mathematical approaches to managing queues, including simulation.
ποΈ Topics Covered
The lecture begins by establishing that waiting lines are non-value-added occurrences that form due to random arrivals, using examples from cricket stadiums and service stations. It defines queuing theory and its goal of minimizing the sum of customer waiting costs and service capacity costs. The lecture then details system characteristics including population source (infinite/finite), number of servers (channels), arrival/service patterns, and queue discipline. It introduces key distributions (Poisson and negative exponential), waiting line models (patient, reneging, jockeying, balking), and the relationship between waiting time and utilization. Finally, it covers simulation as a descriptive technique, including Monte Carlo simulation, computer simulation languages, and the advantages and limitations of simulation.
π Lecture Summary
Waiting Lines
Waiting lines are non-value-added occurrences formed at airports, cricket stadiums, and post offices due to non-scheduled random arrivals. They are often regarded as poor service quality. Examples include orders waiting to be filled, trucks waiting to be loaded/unloaded, jobs waiting to be processed, equipment waiting to be loaded, and machines waiting to be repaired.
Queuing theory is the mathematical approach to the analysis of waiting lines. The goal of queuing analysis is to minimize the sum of two costs: customer waiting costs and service capacity costs. Waiting lines are non-value-added occurrences.
π Definition β Queuing theory: Mathematical approach to the analysis of waiting lines.
Implications of Waiting Lines
Implications include the cost to provide waiting space, loss of business (customers leaving or refusing to wait), loss of goodwill, reduction in customer satisfaction, and congestion that may disrupt other business operations.
Queuing Analysis
Organizations carry out queuing analysis to ensure they can balance service levels with costs. The ultimate goal of queuing analysis is to minimize the sum of two costs: service capacity cost (represented on the x-axis) and customer waiting costs.
Negative Exponential Distribution is another example of a common queuing system. The probability that service time (T) is greater than or equal to a given time (t) is: ( P(T \geq t) = e^{-\mu t} ) where ΞΌ is the average service rate.
π‘ Why this matters: The trade-off between service capacity cost and customer waiting cost is fundamental β increasing capacity reduces waiting but increases costs.
System Characteristics
- Population Source: Infinite source means customer arrivals are unrestricted; finite source means the number of potential customers is limited.
- Number of observers (channels)
- Arrival and service patterns
- Queue discipline (order of service)
Queue discipline is a primary requirement in service systems, but hospital emergency rooms, rush orders in a factory, and mainframe computer processing of jobs do not follow queue discipline.
Elements of Queuing System
Population Source, Arrivals, Waiting Lines, Processing Order, Service, System, and Exit are the common identifiable elements of a Queuing System.
Queuing Systems
The system characteristics are: 1. Population Source; 2. Number of Servers (Channels); 3. Arrival and Service Patterns; 4. Queue Discipline.
A channel is a server in a service system. Systems can have multiple channels and multiple phases.
Poisson Distribution
Poisson distribution is a discrete probability distribution that expresses the probability of a number of events occurring in a fixed period of time if these events occur with a known average rate and are independent of the time since the last event.
Waiting Line Models
- Patient: Customers enter the waiting line and remain until served
- Reneging: Waiting customers grow impatient and leave the line
- Jockeying: Customers may switch to another line
- Balking: Upon arriving, customers decide the line is too long and decide not to enter the line
Waiting Time vs. Utilization
An increase in system utilization comes at the expense of an increase in both the length of the waiting line and average waiting time. These values increase as utilization approaches 100 percent. Under normal circumstances, 100 percent utilization is not a realistic goal.
System Performance
Measures include: 1. Average number of customers waiting; 2. Average time customers wait; 3. System utilization; 4. Implied cost; 5. Probability that an arrival will have to wait
Queuing Models: Infinite-Source
- Single channel, exponential service time
- Single channel, constant service time
- Multiple channel, exponential service time
- Multiple priority service, exponential service time
Priority Model: Arrivals are assigned a priority as they arrive.
Finite-Source Formulas
For a finite population:
| Symbol | Meaning |
|---|---|
| N | Number in population |
| J | Number not waiting or being served |
| L | Number waiting |
| H | Number being served |
| S | Number of servers |
| T | Average service time |
| U | Average time between service requests |
| X | Service factor = T / (T + U) |
| F | Efficiency factor |
| W | Average waiting time |
Key formulas:
- Service Factor: ( X = \frac{T}{T + U} )
- Average Number Waiting: ( L = N(1 - F) )
- Average Waiting Time: ( W = \frac{L(T + U)}{N - L} = \frac{FT(1 - X)}{X} )
- Average Number Being Served: ( H = FNX )
- Average Number Running: ( J = NF(1 - X) )
- Number in Population: ( N = J + L + H )
Formula used: ( F = \frac{H}{J + H} )
Other Approaches - Non-Mathematical
- Reduce perceived waiting time
- Magazines in waiting rooms
- Radio/television
- In-flight movies
- Filling out forms
- Derive benefits from waiting
- Place impulse items near checkout
- Advertise other goods/services
Simulation
Simulation is a descriptive technique that enables a decision maker to evaluate the behavior of a model under various conditions. It models complex situations, models are simple to use and understand, models can play "what if" experiments, and extensive software packages are available.
Simulation Process
- Identify the problem
- Develop the simulation model
- Test the model
- Develop the experiments
- Run the simulation and evaluate results
- Repeat 4 and 5 until results are satisfactory
Monte Carlo Simulation
Monte Carlo method is a probabilistic simulation technique used when a process has a random component. Steps: 1. Identify a probability distribution; 2. Setup intervals of random numbers to match probability distribution; 3. Obtain the random numbers; 4. Interpret the results.
Distributions commonly used are the Poisson and Normal Distributions:
- Poisson distribution: Mean of distribution is required
- Normal Distribution: Need to know the mean and standard deviation
Simulated Value = Mean + Random Number Γ Standard Deviation
Uniform Distribution: Simulated Value = a + (b - a)(Random number as a percentage)
Computer Simulation
Simulation languages include: 1. SIMSCRIPT II.5; 2. GPSS/H; 3. GPSS/PC; 4. RESQ
Advantages of Simulation
- Solves problems that are difficult or impossible to solve mathematically
- Allows experimentation without risk to actual system
- Compresses time to show long-term effects
- Serves as training tool for decision makers
Limitations of Simulation
- Does not produce optimum solution
- Model development may be difficult
- Computer run time may be substantial
- Monte Carlo simulation only applicable to random systems
Why Simulation is Necessary
- Mathematics involved is too complicated
- Easier to manipulate than reality
- Software and hardware permit modeling
Simulation Steps
- Problem formulation
- Model building
- Data acquisition
- Model translation
- Verification & validation
- Experiment planning & execution
- Analysis
- Implementation & documentation
Operations Strategy
The central idea for formulating an Operations Strategy for Waiting Line concept is designing a service system to achieve a balance between service capacity and customer waiting time. The operations strategy should identify an appropriate and acceptable level of service capacity and quality so waiting lines are not formed or are manageable and acceptable to customers. Organizations often engage waiting customers in activities to make the waiting time less painful and more pleasant.
β Key Takeaways
The fundamental goal of queuing analysis is to minimize the total cost which is the sum of service capacity costs and customer waiting costs β these two costs have an inverse relationship. Waiting lines form even in underloaded systems due to random arrival patterns and variability in service times, with system performance measured by average number waiting, average waiting time, system utilization, and implied cost. The Poisson distribution models arrival rates while the negative exponential distribution models service times, and queue discipline (the order of service) is a critical system characteristic. Simulation, particularly Monte Carlo simulation, is essential for analyzing complex queuing systems that cannot be solved mathematically, though it does not produce optimal solutions. The operations strategy for waiting lines should focus on balancing service capacity with customer waiting time, which may include non-mathematical approaches like engaging waiting customers to reduce perceived waiting time.
π§ Quick Revision Questions
- What are the two costs that queuing analysis aims to minimize, and how are they related?
- What is the difference between an infinite source and a finite source population in queuing systems?
- List and briefly define the four waiting line models: patient, reneging, jockeying, and balking.
- What are three advantages and three limitations of using simulation for queuing analysis?
- How do you calculate the service factor (X) in a finite-source queuing model, and what does it represent?