Comprehensive Operations Management Study Notes: Work in Process, Flow Time, and Little's Law

Work in Process and Utilization Calculations

• Formula for Average Work in Process (WIP):     $\text{Average WIP} = \sum (\text{Number of Flow Units}) \times (\text{Fraction of Time Present})$

• Single-Resource Process Example:

  • Process timing over an $80$\text{-second} window:

    • $50\,\text{seconds}$: Exactly $1$ customer present.

    • $30\,\text{seconds}$: Exactly $0$ customers present.

  • Fraction of time with $1$ customer: $\frac{50}{80} = \frac{5}{8}$.

  • Fraction of time with $0$ customers: $\frac{30}{80} = \frac{3}{8}$.

  • Average Work in Process for Resource 1 ($M_1$):         $\text{Average WIP}_{M1} = \left(1 \times \frac{5}{8}\right) + \left(0 \times \frac{3}{8}\right) = 0.625\,\text{customers}$

• Relationship Between WIP and Utilization:

  • When a single resource serves a single flow unit at a time, Average WIP equals Resource Utilization ($0.625$ or $62.5\%$).

  • If batch collection or multi-unit processing occurs (e.g., serving $5$ flow units simultaneously), Average WIP differs from utilization:         $\text{Average WIP}_{\text{batch}} = 5 \times \frac{5}{8} = 3.125\,\text{flow units}$

• Multi-Stage System Work in Process:

  • For a system where $M_1 = 0.625$, $M_2 = 0.625$, and $M_3 = 1.0$:         $\text{Total System WIP} = 0.625 + 0.625 + 1.0 = 2.25\,\text{customers}$

  • For an alternate system configuration where $M_1 = 0.5$, $M_2 = 0.5$, and $M_3 = 0.8$:         $\text{Total System WIP} = 0.5 + 0.5 + 0.8 = 1.8\,\text{customers}$

Actual Cycle Time vs. Flow Time

• Actual Cycle Time (Cycle Time):

  • The time elapsed between consecutive departures of completed flow units from a process.

  • Measures the operational time difference between different flow units exiting the system.

  • Formula:         $\text{Cycle Time} = t_{\text{departure}, n} - t_{\text{departure}, n-1}$

  • Example: If a system completes a unit every $50\,\text{seconds}$, Cycle Time $= 50\,\text{seconds}$.

• Flow Time (Throughput Time / Dwell Time):

  • The total duration spent by a single, specific flow unit inside process boundaries from the moment it enters to the moment it leaves.

  • Tracks the exact same flow unit across its entire journey.

  • Formula:         $\text{Flow Time} = t_{\text{departure, unit } A} - t_{\text{arrival, unit } A}$

• Stakeholder Perspectives:

  • Manager Perspective: Focuses primarily on Cycle Time because it reflects process speed, capacity, and service delivery rates.

  • Customer Perspective: Focuses primarily on Flow Time (e.g., a customer at Chipotle or a service center cares about their total personal elapsed time in line and service, not the rate at which other customers leave).

• Assembly Line Demonstration (4-Stage Toy Car Assembly):

  • Setup: 4 sequential assembly stages, each requiring $1\,\text{minute}$ of processing time.

  • Process Progression:

    • $t = 0\,\text{min}$: Car 1 enters Stage 1.

    • $t = 1\,\text{min}$: Car 1 advances to Stage 2; Car 2 enters Stage 1.

    • $t = 2\,\text{min}$: Car 1 advances to Stage 3; Car 2 moves to Stage 2; Car 3 enters Stage 1.

    • $t = 3\,\text{min}$: Car 1 advances to Stage 4; Car 2 moves to Stage 3; Car 3 moves to Stage 2; Car 4 enters Stage 1.

    • $t = 4\,\text{min}$: Car 1 exits Stage 4 and leaves the process.

    • $t = 5\,\text{min}$: Car 2 exits Stage 4 and leaves the process.

  • Analysis of Metrics:

    • Flow Time for Car 1 $= 4\,\text{minutes}$ (tracked from entry at $t=0$ to exit at $t=4$).

    • Cycle Time $= 1\,\text{minute}$ (time difference between departure of Car 1 at $t=4$ and Car 2 at $t=5$).

Flow Time with Buffers and Alternating Patterns

• Inclusion of Waiting Time:

  • Flow time is not restricted to active processing time; it includes all queueing, storage, and buffer delays.         $\text{Flow Time} = \text{Active Processing Time} + \text{Waiting Time}$

• Comparative Benchmark Examples:

  • Subway Sandwich Shop:

    • Cycle Time $= 30\,\text{seconds}$ (a customer exits every $30\,\text{seconds}$).

    • Flow Time $= 3\,\text{minutes}$ (a customer spends $3\,\text{minutes}$ on average inside the store).

  • Automobile Assembly Line:

    • Cycle Time $= 1.5\,\text{minutes}$ (a completed car rolls off the line every $1.5\,\text{minutes}$).

    • Flow Time $= 90\,\text{minutes}$ (a single vehicle spends $90\,\text{minutes}$ total in production).

• Calculating Flow Time via Arrival and Departure Tracking:

  • Customer 1: Enters at $t = 0\,\text{seconds}$, departs at $t = 90\,\text{seconds}$.         $\text{Flow Time}_1 = 90 - 0 = 90\,\text{seconds}$

  • Customer 2: Enters at $t = 0\,\text{seconds}$, experiences intermediate buffer waiting ($50\,\text{seconds}$ waiting + $40\,\text{seconds}$ stage 1 + $40\,\text{seconds}$ stage 2), departs at $t = 130\,\text{seconds}$.         $\text{Flow Time}_2 = 130 - 0 = 130\,\text{seconds}$

• Averaging Alternating Flow Time Patterns:

  • When process data reveals repeating cycles of varying flow times (e.g., $90\,\text{s}, 130\,\text{s}, 90\,\text{s}, 130\,\text{s}$):

    1. Identify the base repeating unit/pattern.

    2. Compute the average across the numbers in that repeating pattern.

  • Calculation for pattern $(90\,\text{s}, 130\,\text{s})$:         $\text{Average Flow Time} = \frac{90 + 130}{2} = 110\,\text{seconds}$

  • Calculation if pattern expands to $(90\,\text{s}, 130\,\text{s}, 150\,\text{s})$:         $\text{Average Flow Time} = \frac{90 + 130 + 150}{3} = 123.33\,\text{seconds}$

Little's Law and Key Operations Metrics

• Definition and Origin:

  • Proposed by MIT Professor John Little.

  • Establishes the fundamental operational relationship between Work in Process ($\text{WIP}$), Flow Rate ($R$), and Flow Time ($T$).

• Mathematical Formula:     $\text{Work in Process } (\text{WIP}) = \text{Flow Rate } (R) \times \text{Flow Time } (T)$     $\text{WIP} = R \times T$

• Operational Significance:

  • Eliminates the need to construct complex Gantt charts to evaluate long-term process inventory or timing.

  • Allows any single variable to be derived if the other two are known:         $R = \frac{\text{WIP}}{T}$         $T = \frac{\text{WIP}}{R}$

• Deriving Flow Rate from Cycle Time:

  • Flow Rate is the reciprocal of Cycle Time:         $\text{Flow Rate } (R) = \frac{1}{\text{Cycle Time}}$

  • Example: If Cycle Time $= 40\,\text{seconds}$, then:         $R = \frac{1}{40\,\text{seconds}} = 0.025\,\text{units/second} = 90\,\text{units/hour}$

Core Operational Definitions and Units

• Work in Process (WIP):

  • Definition: The total number of flow units contained within the process boundaries at any given instant (e.g., customers in queue/service, car chassis in assembly, active claims in processing).

  • Unit of Measurement: Discrete units or scalar counts (e.g., $\text{cars}$, $\text{customers}$, $\text{claims}$).

• Flow Time (T):

  • Definition: The total elapsed time spent by a flow unit inside process boundaries, including active work duration and queueing delays.

  • Unit of Measurement: Time units (e.g., $\text{seconds}$, $\text{minutes}$, $\text{hours}$, $\text{days}$, $\text{years}$).

• Flow Rate (R):

  • Definition: The rate at which the process delivers completed output over a specific time window; represents throughput speed.

  • System Capacity Formula:         $\text{Flow Rate } (R) = \min(\text{Process Capacity}, \text{Demand Rate})$

  • Unit of Measurement: Units per unit of time (e.g., $\text{customers/hour}$, $\text{cars/month}$, $\text{claims/week}$).

• Dimensional Consistency Rule:

  • When multiplying Flow Rate ($R$) and Flow Time ($T$) to calculate WIP, the time units of both parameters must be identical so that time units cancel out, leaving pure unit counts.

  • Example: If Flow Rate $R = 4\,\text{cars/minute}$ and Flow Time $T = 1\,\text{minute}$ ($60\,\text{seconds}$):         $\text{WIP} = (4\,\text{cars/minute}) \times (1\,\text{minute}) = 4\,\text{cars}$

Numerical Applications and Problem Sets

• Application 1: University Student Graduation Time:

  • Flow Unit: Student.

  • Flow Rate ($R$): Average admitted/graduating class size per year $= 1,373\,\text{students/year}$.

  • Work in Process (WIP): Total enrolled student population active at the university.

  • Flow Time ($T$): Average graduation time (years required to graduate).

  • Formula Setup:         $\text{Average Graduation Time } (T) = \frac{\text{Total Enrolled Students } (\text{WIP})}{\text{Average Class Size } (R)}$

  • Calculation Example:

    • If total enrollment $\text{WIP} = 8,238\,\text{students}$:             $T = \frac{8,238\,\text{students}}{1,373\,\text{students/year}} \approx 6\,\text{years}$

• Application 2: Insurance Claim Processing (Burpee Problem):

  • Given Data:

    • Annual claim volume: $6,000\,\text{claims/year}$.

    • Processing flow time per claim ($T$): $2\,\text{weeks}$.

  • Unit Conversion:

    • Convert annual flow rate to weekly flow rate using $52\,\text{weeks/year}$:             $R = \frac{6,000\,\text{claims}}{52\,\text{weeks}} = 115.385\,\text{claims/week}$

  • Work in Process Calculation:         $\text{WIP} = R \times T = 115.385\,\text{claims/week} \times 2\,\text{weeks} \approx 230.77\,\text{claims}$

• Application 3: Gas Station / Service Stop Calculations:

  • Given Data:

    • Work in Process ($\text{WIP}$): $2\,\text{units}$ (or active stops).

    • Flow Time ($T$): $30\,\text{seconds}$.

  • Hourly Flow Rate Calculation:

    • Convert $30\,\text{seconds}$ to hours:             $30\,\text{seconds} = \frac{30}{3,600}\,\text{hours} = \frac{1}{120}\,\text{hours}$

    • Apply Little's Law to find hourly flow rate:             $\text{Flow Rate } (R) = \frac{\text{WIP}}{T} = \frac{2}{\frac{1}{120}\,\text{hours}} = 240\,\text{stops/hour}$

  • Annual Throughput Calculation:

    • For a process delivering $220\,\text{stops/hour}$ up to an annual peak rate of $453.08\,\text{stops/hour}$ over $8,760\,\text{hours/year}$:             $\text{Annual Flow Rate} = 453.08\,\text{stops/hour} \times 8,760\,\text{hours/year} = 3,969,000\,\text{stops/year}$

Audience Questions and Discussion

• Question: Does average work in progress always match resource utilization?

  • Response: Average WIP matches utilization only when there is exactly one resource serving one flow unit at a time. If there is batch collection or multi-unit processing (e.g., serving 5 flow units at a time), WIP differs and is multiplied by the batch size ($5 \times \text{utilization}$).

• Question: Can flow time be determined simply by adding individual station activity times (such as summing $M_1$ and $M_3$)?

  • Response: Summing activity times gives total flow time only if there are zero buffer delays or queues in the system. If buffers exist, waiting time must be added to activity times. Measuring departure time minus arrival time ($\text{Departure} - \text{Arrival}$) guarantees accurate inclusion of all delays.

• Question: What unit format results from multiplying flow time and flow rate in Little's Law?

  • Response: Time units in the flow rate denominator and flow time numerator cancel out completely (e.g., $\text{weeks} \times \frac{\text{claims}}{\text{week}} = \text{claims}$), resulting in a discrete unit count representing Work in Process.