Which are the Inputs/outputs of the Production scheduling table?
Input= Production plan. Output= Plan of jobs, Plan of people, Plan of resources.
Mention some Critical issues in Production scheduling.
*Huge amount of data:
- technical
- managerial
- Uncertainty of working data -variability (e.g product variants, etc)
- unpredictability (eg. failures,etc))
- Uncertainity (eg. rowking times, etc)
- Difficulties in the objective definition.
- efficiency (e.g machine saturation)
- or effectiveness (delay minimization, low completion times)
In Production Scheduling, define Due Date and mention others plans related with Due Date jobs.
Due Date (DD): It is a date in which a job should be completed (forced by higher planning levels or by final client)
- EDD = Earliest Due Date: The earliest planned date an order may be completed
- LDD = Latest Due Date: The latest planned date an order may be completed
- ESD = Earliest Start Date: The earliest planned date at which it is possible to start working the order
- LSD = Latest Start Date: The latest planned date at which it is possible to start working the order without missing its due date
- Slack = DD – Processing time – ESD (static, before releasing job)
In Production Scheduling, which is the meaning of JOB?
- It is a working unit composed by physical units that are all of the same kind.
It could be a single piece or a box or a batch.
It could happen that there is no correspondence between job and client order.
In Production Scheduling, which is the meaning of ORDER?
The product or service required by a customer.
In Production Scheduling, which is the meaning of ROUTING?
-The sequence of stages of production that an order has to go through. The routing specifies many details including standard time, tooling required, etc.
In Production Scheduling, which is the meaning of PREEMPTION?
-Stopping of the processing of one job on a machine in order to process another one that is more urgent.
In Production Scheduling, which is the meaning of PASSING?
- Overtaking between jobs waiting to be processes, for urgency reasons.
Which are the Classifications profiles in Production Scheduling? (be schematic in the 3-axis diagram)
X - Specific objectives.
Y - Production system typology
- plant configuration
- type of constraints considered
Z - Used resolution technique.
Mention some of the “possible production system configurations” in Production Scheduling.
- Single machine
- Parallel machine
- Job Shop
- Flow Shop
- Open Shop
About the basic scheduling principles.. What does the FORWARD SCHEDULING consist?
-It starts from the ESD (static, before realising job) of each job and plan ahead the individual operations of the routing of the job, calculating, with the help of the lead time of each operation, the ESD and EDD of each operation.
The processing is done for all jobs, considering operations already allocated and, thus, also the EDD (Earliest Due Date) of the plan is calculated (the latest of the EDD’s).
About the basic scheduling principles.. What does the BACKWARD SCHEDULING consist?
-It starts from the DD (Due Date) of each job (equal to the LDD (Latest Due Date) ) and individual operations of the job are planned BACKWARDS calculating their LDD and LSD (Latest Start Date), based on their routing and related lead times of each operation. The Schedule is done for all jobs and, thus, also the LSD of the plan is obtained (the smallest of the LSD)
The operations are allocated on the various resources of the system defined by the technological routing, also considering the operations allocated by other jobs.
*Mention the main objectives in Basic Scheduling
- Minimize average Lateness.
- Minimizing average tardiness.
- Minimizing average Flowtime.
- Minimizing number of late jobs (tardiness different from zero)
- Minimizing “Makespan”
- Maximizing average rate of saturation of the system (working time / available time).
- Composed Weighted Criteria of the previous ones.
How can we define the Priority Index?
The job to be loaded on the machine is chosen among the ones available based on a priority index (also referred as a “Figure of worthiness”).
-The priority index is computed for each job based on one or more of the following variables:
- Processing time
- Set-up time
- Due date
- Job state
- Machine/plant state
- Delay costs
How can we define the Dispatching Rules?
- Often, due to the COMPLEXITY and the DYNAMIC NATURE of the SYSTEM and ENVIRONMENT analyzed, an operative plan, that pre-defines allocations and operations sequences, is not formulated: if it’s formulated, it happens to be interrupted and often cancelled, so not generating good results
- Instead, RULES to take decisions each moment are provided
- This approach is generally (with some rare exceptions) HEURISTIC but flexible and relatively easy to use
Provide the scheme of the Dispatching Rules (3-axis diagram)
X - Information Update
- Static:
The reciprocal priority of two jobs in queue at the same machine does not change over time
The priority index of each job can be computed once for all when the job enters the queue at the machine
- Dynamic:
The reciprocal priority of two jobs in queue at the same machine can change over time
The priority index of each job must be re-computed any time a decision is taken
Y - Variables used for making decision
Z - Information Extent
- local: The priority index for a job on a machine is calculated only with information regarding the operations on that (i.e. only the next) machine
- Global: The priority index for a job on a machine is calculated with information regarding also the operations on other (i.e. all the remaining) machines
Provide examples of Dispatching Rule
*Processing time or Set-up time
-SPT (Shortest Processing Time) [see later]
-LPT (Longest Processing Time)
-TSPT (Truncated Shortest Processing Time): as SPT, but with a waiting time threshold; when overcome, the FIFO [see later] is applied
-LWKR (Least Work Remaining): min. residual total processing time
-TWORK (Total Work): min. total processing time
-MSUT (Minimun Set-up Time): min. set-up time on the machine
…
*Due date
-EDD (Earliest Due Date) [see later]
-OPNDD (Operation Due Date) [see later]
- System status
- NINQ (Number In Next Queue): min. number of jobs already queued for the next operation
- WINQ (Work In Next Queue): min. total processing time (i.e., sum of the processing time of jobs) already queued for the next operation
*Job status
FIFO (First In First Out)
LIFO (Last In First Out)
FISFS (First In the System First Served): the first job that entered the system for the first time
FROP (Fewest Remaining Operations): min. number of residual operations
MROP (Most Remaining Operations): max. number of residual operations
*Economic factors
COVERT: max. ratio = delay cost / residual time
*Weighted rules
SPT/LWKR: min PT + (1- ) RT (PT = processing time; RT = residual total processing time)
Mention the Benefits of Dispatching Rules
- Simple
- Limited number of information needed
- Flexibility of use
- Easy to transfer and to reconfigure
- They are not strictly linked to production system characteristics
- Decision taken at the latest
- A plan defined a priori becomes obsolete after few hours
- Good practical results
- Few valid alternatives for a job shop
In which constist The Shortest Processing Time. [SPT]
- Variable: processing time
- Static
- Local
- First the job with the shortest processing time on the considered machine
- In case of single machine, it minimizes the average flowtime
- In general, it provides very good results in terms of average flowtime
- Risk of large delays of the longest jobs
In which consist The Earliest Due Date [EDD]
- Variable: Due date
- Static
- Local
- First the job with the earliest due date
- It provides very good results in terms of number of tardy jobs
- It neglects the processing time therefore it penalizes the longest jobs
In which consist Slack per Operation [S/OPN]]
- Variables: Due date, processing time, present time (now)
- Dynamic
- Global
- Compute for each job the SLACK = due date – present time – residual total processing time
- Compute the rate S/OPN = SLACK / number of residual operations
- First the job with lowest S/OPN
- It provides very good results in terms of punctuality (especially if the due date have been “reasonably” defined)
- It needs to replicate a computation every time the rule is applied
Define a Local Index of Information Extent
-Priority Index for a job on a machine is calculated only with information regarding the operations on that machine. (i.e. only the next)
Define a Global Index of Information Extent
The priority index for a job on a machine is calculated with information regarding also the operations on other machines (i.e. all the remaining machines).
Define in the Information Update the Static Index.
- The reciprocal priority of two jobs in queue at the same machine does not change over time.
- The priority index of each job can be computed once for all when the job enters the queue at the machine.
