Chapter 5: Capacity Planning Flashcards
Capacity
The maximum sustainable flow-rate of output of a process or a system
determines possible throughput
Long term capacity planning considerations
Economies and diseconomies of scale
Capacity timing and sizing
Capacity cushions
Trade off between customer service and capacity utilization
Short term constraint management considerations
- theory of constraints
- identification and management of bottlenecks
- product mix decisions using bottlenecks
-managing constraint in a line process
Output measures of capacity
Capacity measured in terms of outputs. Best utilized for individual processes within a firm or when a firm provides a relatively small number of standardize services/ products
Becomes less useful as variety in product mix increases
Input measures of capacity
Generally used for low volume, flexible processes
Issue is that demand is generally expressed as an output rate and must be converted
Utilization
Aka capacity utilization
The degree to which a resource is currently being used
= Average output rate / max capacity (expressed as a percent)
(= throughput / theoretical capacity)
Output rate and capacity must be measured in the same terms
Maximum capacity for the utilization equation
Greatest level of output that a process can reasonably sustain for a longer period using realistic schedules and current equipment
Exceeding maximum capacity can be done for peaks but cannot be sustained
Economies of scale
The average unit cost of a good or service can be reduced by increasing it’s output rate because:
- spread out fixed costs
- reducing construction costs
- cutting costs of purchased materials (volume discounts)
- finding process advantages (line processes) (can afford to dedicate resources)
Diseconomies of scale
Occur when the average cost per unit (at the best operating level) increases as a facility’s size increases
Size can bring complexity, loss of focus, inefficiency
Three dimensions of capacity strategy
- sizing capacity cushions
- timing and sizing expansion
- linking process capacity and other operating decisions
Capacity cushions
Provide a buffer against uncertainty
Amount of reserve capacity a process uses to handle sudden increase in demand or temporary losses in production capacity.
= 100% - average utilization rate %
Size of capacity cushion
Varys by industry
Often smaller in capital intensive industries and larger where service time is paramount.
Large cushions important when demand varies or future demand is uncertain and resource flexibility is low or if prompt customer service is a competitive priority
But unused capacity costs money
Small cushions may uncover inefficiencies, usually implies less risk and less idle capacity
Expansionist strategy for expanding capacity
Large, infrequent jumps in capacity. Attempts to stay ahead of demand and minimize loss to insufficient capacity
aka capacity lead strategy
Wait-and-see strategy for expanding capacity
Smaller, more frequent jumps in capacity. Focus on short term options for increases: lags behind demand
Reduced risk for overexpansion/ obsolete technology but risks being preempted by a competitor
aka capacity lag strategy
Four step procedure for making capacity decisions
- Estimate future capacity requirement
- Identify gaps by comparing requirements with available capacity
- Develop alternative plans for reducing gaps
- Evaluate each alternative qualitatively and quantitatively before making a final choice
Capacity requirement
What a process’s capacity should be for some future time period to meet the demand of customers (internal or external) given the firm’s desired capacity cushion
Can be measured by output or input
Planning horizon
The set of consecutive time periods considered for planning purposes
When input measures are appropriate
- Product variety and process divergence is high.
- product or service mix is changing
- productivity rates are expected to change
- significantly learning effects are expected
Capacity requirement equation
For a single service or process, with a time period of a year
= (Processing hours required for years demand) / (hours available from a single capacity unit after deducting the desired capacity cushion)
Alternate capacity requirement equation
Number of input units required = (demand forecast * processing time) / (total number of hours available in a year from one unit of capacity * (1-capacity cushion as a decimal))
Setup time
The time required to change a process or operation from making one service or product to another
Equation for setups per year
Number of set ups per year= (Number of units forecast per year / number of units made in each lot requiring setup)
Equation for total setup time
= number of setups per year * time per setup
Capacity requirement numerator accounting for setup time)
For each product
(Demand forecast for year * processing time) + ((demand forecast for the year/ units in each lot)* time per setup)
Can add as many of these as needed together in the numerator to capture all services/ processes
Capacity gap
Positive or negative difference between projected demand and current capacity
Base case
In making capacity decisions: The act of doing nothing and losing orders from any demand that exceeds current capacity or incur costs because capacity is too late
All alternative plans compared to this
Qualitative consideration for alternative capacity plans
Uncertainty of demand
Competitive reaction
Technological change
What-if analysis
Major when entering new markets/ changing strategy
Cash flow
The difference between the flows of funds into and out of an organization over a period of time (including changes in assets and liabilities)
Quantitative assessment of capacity alternative
Estimate cash flows attributable to each alternative
Integrated resource planning
IRP
Long-term capacity plan for utilities to meet the growing forecasted annual energy demand
Capacity cushions for peak times
Waiting line models
Use probability distributions to provide estimates of average customer wait time, average length of waiting lines, and utilization of the work center. Can be used to choose cost effective capacity (gains from increasing service efficiency against costs to do so)
Demand trees
Good for uncertain demand and sequential decisions
Can include additional cost incurred or financial benefits expected at each decision point
Waiting line model: single-server
Single-channel, single-phase (one server, one line
Waiting line model variables
λ: mean arrival rate of customers (inventory). A poisson distribution
μ: mean service rate. Exponential distribution
ρ: average utilization of the system
Assuming mean service rate > mean arrival rate
Probability that n customers are in the system (waiting line model)
P(sub n): probability that n customers are in the system = (1-ρ)ρ^n
P(sub 0): probability that 0 customers are in the system = 1-ρ
Average number of customers in the service system (waiting line model)
L = λ / (μ - λ)
Mean arrival rate / (mean service rate - mean arrival rate)
Average number of customers waiting in line (waiting line model)
L (sub q) = ρL
Average utilization of the system * average number of customers in the service system
Average time spent in the system including service (waiting line model)
W = 1 / (μ - λ)
1 / (mean service rate - mean arrival rate)
Average waiting time in line (waiting line model)
W(sub q) = ρW
Average utilization do the system * average time spent in the system including service
Multiple server model (waiting line model)
Customers (inventory) form a single line and choose/ go to one of s servers when one is available
Multiple channel, single phase
Additional variables for multiple server model
s = identical servers
Service distribution is exponential
1/μ mean service time
sμ exceeds λ (arrival rate)
Littles law
L = λW
Average number of customers in the service system = arrival rate * average time spent in system
Can always estimate the third variable if have any given two
aka inventory = flow rate * flow time
Finite source waiting line model
Used if customer population is finite, less than 30
Process elements that determine capacity
Nature and mix of flow units (more variety = slower processes)
- required activities and buffers
- resources allocated to processing flow units
-operating procedures used to manage actvities
Resource pool
collection of interchangeable resources that can perform an identical set of activities
each unit is called a resource unit
Theoretical capacity of a resource pool
maximum number of flow units that can be processed by the pool per unit of time the pool may be fully utilized
can be computed for the entire process/ resource pool or for each unit (this assumes no waiting or delays)
1/unit load* load batch * scheduled availability
(unit load = time per batch)
Bottleneck resources
resources pools in a process with the minimum capacity - provide a limit on the theoretical process capacity
to increase overall capacity must increase bottleneck capacity
Unit load of a resource unit
the sum of the work contents of all activities that utilize that resource unit
measured in units of time required per flow unit
Units per period or loads per period
Load batching
when a resource unit can process several flow unit simultaneously
load batch
the number of units a resource unit can process at once
scheduled availability of a resource pool
the sum of the scheduled availability of all units in the pool
Factors affecting capacity utilization
resource breakdown
preventative maintenance
setup and breakdown (resource unavailable)
low demand or supply
starvation/ blockage
Potential ways to improve capacity
- increase net availability of resources (reduced set up times, technological improvements, better sequencing, planned maintenance
- synchronize flows to decrease idleness
- increase theoretical capacity (address bottlenecks)
Interdependent product lines
distinct marketing products with some production activities that are integrated/ similar/ symbiotic (using the same bases/ materials/ resources)
may enable economies of scale
average capacity strategy
add capacity to coincide with AVERAGE growth in demand
Techniques to increase short-run capacity
(because forecasts are imperfect)
- increase resources
- improve resource usage
- modify the output
- modify the demand
is actually easier to increase short run capacity than decrease. generally a decrease in demand causes capacity to go unused
