For optimizing an RMG (Ready-Made Garment) in-house facility, focusing on objectives such

as reducing lead time, increasing worker efficiency, and decreasing work-in-progress

(WIP) inventory, the number of data points you’ll need depends on the complexity of the
problem and the decision variables involved. Here’s an estimate based on your specific
objectives:

1. Key Decision Variables for RMG Facility Optimization

(a) Lead Time Reduction

● Worker Performance: Time taken for each worker to complete their task (e.g., sewing,
cutting).
● Machine Utilization: Time machines are in use versus idle times, downtime for
maintenance, etc.
● Job Sequencing: Order in which tasks are assigned to workers or machines.
● Shift Scheduling: Worker shift patterns and overlaps.
● Batch Size: Size of production batches for each stage of manufacturing.
● Material Handling Time: Time spent moving materials between different stages.

(b) Worker Efficiency Improvement

● Cycle Time per Operation: Time workers spend on each task.

● Worker Skills: Skill levels and the effect on task completion times.
● Task Assignment: How tasks are distributed among workers to maximize skill use.
● Training Time: Time allocated to training workers to improve their speed/efficiency.
● Worker Fatigue: Impact of working hours on productivity (e.g., peak vs. low productivity
times).

(c) Decreasing WIP Inventory

● WIP Inventory Levels: Number of items at each stage of production.

● Rework Rates: Percentage of items that require rework, increasing WIP.
● Production Flow: Smoothness and bottlenecks in the flow of goods through stages.
● Material Arrival: Timing and frequency of material deliveries impacting inventory
buildup.

2. Estimating Data Points for Each Objective

(a) Lead Time Reduction:

● Key decision variables: Worker performance, job sequencing, batch size, machine
utilization, material handling times.
 ● Estimated data points: At least 100-500 data points across various operations,
machines, and production lines. You’ll want enough data to account for variability in
production cycles, machine downtimes, and worker productivity.

(b) Worker Efficiency Improvement:

● Key decision variables: Worker task cycle time, skill levels, task assignment, worker
fatigue, training time.
● Estimated data points: To capture different productivity levels at various stages, 200-600
data points. Each worker’s performance needs to be tracked over multiple time periods
to capture how training, fatigue, or shift patterns affect their efficiency.

(c) Decreasing WIP Inventory:

● Key decision variables: Inventory levels at different stages, rework rates, production flow
bottlenecks, material arrival times.
● Estimated data points: To monitor inventory and detect trends over time, you may need
100-400 data points, covering different production runs, shifts, and material arrivals.

3. Total Number of Data Points

For an in-house RMG facility optimization covering all three objectives:

● Lead time reduction: 100-500 data points.

● Worker efficiency: 200-600 data points.
● WIP inventory reduction: 100-400 data points.

Based on these objectives, the total data points you would need for the algorithms to work
effectively will likely be in the range of 400 to 1,500 data points. This accounts for variability
across different production cycles, workers, machines, and operational conditions.

4. Factors Affecting Data Requirements

● Complexity of Operations: The more complex your production process (e.g., different
products, many stages), the more data you'll need to model the entire system effectively.
● Number of Workers and Machines: Larger facilities with many workers and machines
will require more data to capture all interactions and dependencies.
● Level of Granularity: Finer data granularity (e.g., cycle times for individual tasks) will
increase data point needs.

5. Example Data Sources

● Production Logs: Historical data on worker task completion times, machine usage, and
job sequencing.
● WIP Inventory Records: Data on items at various stages in the production line.
 ● Worker Performance: Tracking worker output, training schedules, and fatigue over time.
● Shift Records: Data on worker shifts, machine downtimes, and production line flow.

Data that should be collected

1. Core Operational Data (In-House Facility Optimization)

(a) Lead Time Reduction

● Worker Performance Metrics:

○ Task completion times for each worker (cutting, stitching, etc.).
○ Output per worker per shift.
○ Downtime due to worker breaks or fatigue.
● Machine Utilization:
○ Machine operating times and downtime.
○ Maintenance schedules and breakdowns.
● Job Sequencing Data:
○ Order of tasks assigned to workers or machines.
○ Time spent waiting between task handovers.
● Material Handling:
○ Time spent moving materials between production stages.
○ Delays in material transfer due to bottlenecks.
● Batch Size and Production Rate:
○ Production batch sizes.
○ Average production rate per batch.
● Shift Schedules:
○ Worker shift patterns, overlaps, and productivity at different times.

(b) Worker Efficiency Improvement

● Worker Cycle Times:

○ Time taken by workers to complete specific tasks.
○ Differences in cycle time based on skill levels.
● Task Assignment:
○ How tasks are distributed based on skill or experience.
○ Effectiveness of task allocation on overall efficiency.
● Training Data:
○ Time spent on training workers for new tasks.
○ Efficiency improvement after training sessions.
● Worker Fatigue Data:
○ Performance variation over the length of a shift (e.g., first vs. last hours).
● Absenteeism Rates:
○ Worker absenteeism and its effect on production.
(c) WIP Inventory Reduction

● Inventory Levels:
○ Number of items in WIP at each stage of production.
○ Stock levels at different times during the day.
● Rework Data:
○ Percentage of products requiring rework at each stage.
○ Causes of defects that lead to rework.
● Production Flow:
○ Smoothness of production flow (items per hour moving from one stage to
another).
○ Bottlenecks or delays between stages.
● Material Delivery Data:
○ Timing and frequency of material deliveries.
○ Delays in material arrival.

2. Data for Supply Chain Resilience (Incorporating Variability)

(a) Demand Variability

● Historical Demand Data:

○ Variations in demand for RMG products across different seasons and regions.
○ Frequency of demand spikes and drops.
● Forecast Accuracy:
○ Historical accuracy of demand forecasts.
○ Errors in forecasting that led to over- or under-production.

(b) Supplier and Material Variability

● Supplier Lead Times:

○ Variability in lead times from fabric or accessory suppliers.
○ Impact of supplier reliability on production schedules.
● Material Quality Variations:
○ Changes in material quality and their effect on production (e.g., rejection rates).
○ Frequency of receiving defective or delayed material batches.

(c) Transportation and Logistics Data

● Transit Time Variability:

○ Variability in transportation times for raw materials.
○ Delays caused by transportation issues (e.g., strikes, weather events).
● Logistics Disruptions:
○ Frequency of disruptions such as vehicle breakdowns or route closures.
○ Impact of delays on production schedules.
(d) Worker Availability Variability

● Absenteeism Trends:
○ Variations in worker absenteeism rates due to external factors (e.g., strikes,
illness, political unrest).
● Worker Turnover:
○ Frequency of worker turnover and its impact on efficiency.
● Labor Strikes and Political Instability:
○ Data on past strikes or political events affecting the workforce.

(e) External Disruptions

● Environmental Data:
○ Impact of weather on production and supply chain performance (e.g., floods,
heatwaves).
● Economic Factors:
○ Exchange rate fluctuations, raw material price volatility.
○ Impact of economic downturns on material costs or worker wages.

Probable number of Data points required

Lead Time and Worker Efficiency (100-500 Data Points):

● Worker Performance: To capture variability in how individual workers perform tasks

over time, you would need several data points for each worker over multiple production
cycles. For example:
○ If you have 50 workers, and you want to track their task completion times over 10
different days or production cycles, this would give you 500 data points.
● Machine Utilization: The same logic applies for machine utilization. If you have 10
machines, and you monitor their uptime and downtime across 10 production cycles, that
gives you 100 data points.
● Shift Schedules: Worker shifts need to be analyzed to optimize staffing, scheduling,
and productivity. The more workers and shifts you track over time, the more data points
you generate.

Why this range (100-500)?

● 100 data points is a reasonable minimum to capture the basic variability of smaller
teams or less complex operations. It gives you enough variability to begin optimization,
especially for smaller RMG operations.
● 500 data points or more would be necessary for more complex operations (more
workers, machines, and shifts). This higher number ensures that statistical analysis is
robust enough to detect patterns and anomalies across different production
environments.
2. Inventory Management (100-300 Data Points):

● WIP Levels: Work-in-progress inventory typically fluctuates across different stages of

production. You'd need to collect data on WIP levels multiple times a day, across several
production lines and different stages (cutting, sewing, finishing, etc.).
● Rework Rates: If you track rework rates (the percentage of products that need to be
corrected or redone), you’d gather data on rework per shift or per production run.
● Production Flow: Monitoring the speed of production across different products
(especially if you're producing different types of garments) over multiple cycles will help
you understand where bottlenecks are occurring.

Why this range (100-300)?

● 100 data points might be sufficient if you're tracking just a few products or a smaller
production line.
● 300 data points would be more appropriate for facilities producing multiple garment
types, especially with a higher number of production stages, as you'd need enough data
to make comparisons between products or shifts.

3. Variability and Disruptions (100-500 Data Points):

● Supplier Lead Time Variability: You need to gather data on supplier lead times over a
long period to understand fluctuations and predict future disruptions. This includes
tracking late deliveries or changes in material availability.
● Demand Fluctuations: Demand for RMG products can vary by season, region, and
market. Tracking these fluctuations over time allows you to build robust demand
forecasting models.
● Worker Absenteeism: Absenteeism data will help you optimize shift schedules and plan
for disruptions. The more workers you track, the more variability you capture.
● Transportation Disruptions: Delays in transportation, caused by issues like vehicle
breakdowns, strikes, or route closures, are unpredictable and can significantly affect
your supply chain. Data collection over several months (or across different regions) will
provide insights into the likelihood and impact of such disruptions.

Why this range (100-500)?

● 100 data points would capture essential variability in smaller or simpler supply chains
(fewer suppliers, lower demand variability).
● 500 data points allows you to build more complex models that incorporate multiple
disruptions and variability factors across a larger network of suppliers, markets, and
transportation routes.