Flashcards
8. Queues and Queuing Systems
Flashcards do conjunto (32)
- 1Where queues appear⎝ Queues appear in services, logistics and production (scheduling) but also in transportation, information systems and telecommunication systems. They are linked to competitive strategy. Agner Krarup Erlang, a Danish engineer, published the first paper on queuing theory in 1909.
- 2The structure of queuing systemsInput: Customers, parts, vehicles, machines, jobs, etc.
Queues are not necessarily physical. They can be customers on hold, jobs to be printed or even airplanes circling the airport.
If waiting times are long, then the input might divert to another service provider or a substitute product or service may be in place.
If waiting times are short, then the availability of resources should be checked for reduced efficiency. - 3Queues and their link to competitive strategy• Queues are related for the key processes for service, e.g. interactions with the customer (before, during, after), the transfer of information, the role of the ‘front office’ or ‘contact person’ and the transparency and reliability of information. These particularly relate to queuing on the basis of the management of queues and the facilities for waiting lines.
• Queues are essential to service management, where consumption equals production and where interaction with staff and customers lie at the heart of delivery. Furthermore, the service level dominates capacity planning, so businesses should try to avoid queuing / waiting times, which leads to higher degree of available capacity and lower utilization rates on average. - 4List types of queuing systemsSingle channel
Multiple channel - 5Define single channel queuing systemsSingle channel queuing systems guarantee fairness and there is no customer anxiety regarding choice of queue. This avoids any ‘cutting in’ problems or jockeying (line switching). Single channel is the most efficient set-up for minimizing time in the queue.
- 6Define multiple channel queuing systemsMultiple channel queuing systems require differentiation in service (e.g. express lanes) and perhaps also labor specialization. This means that the customer has more flexibility and that balking behavior may be deterred, since several medium-length lines can be less intimidating than one very long line.
- 7What happens if the queue is too long?If the queue is too long, balking behavior might occur since waiting time is longer. Long waiting time may also lead to ‘cutting in’ and impatience of customers. The capacity might not be able to hold everyone. The stress may impact the capability of serves, indirectly determining capacity.
- 8Explain the priorities rule (queue discipline)1. First come, first served
2. Last in, first out
3. Served in random order
Additional prioritizing:
1. Shortest processing time
2. Reservations first
3. Emergencies first
4. Limited needs
5. (Other rules for setting priorities)
Except for the first, all other modes of setting priorities require some sort of identification (coding) - 9Multiple channels and different capabilities
- 10Multiple channels + multiple stages
- 11The purpose of modelling queuing systems• Modelling of queuing systems integrate the mode of arrivals and the processing by ‘servers’ to determine queue lengths, waiting times and resource allocation. This is done using Little’s Law based on averages.
• Queuing models integrate the mode of arrivals and processing by the ‘servers’ to determine queue lengths, waiting times (for example, Little’s law) and resource allocation. - 12Typical questions for queuing systemsTypical questions for queuing systems would be:
What is the average number of customers in the system?
What is the average time a customer spends in the system?
What is the probability that a customer is rejected?
For what fraction of time is the ‘server’ idle? - 13Modelling of queuing systems: Calculations• Calculations and simulations are approached from a ‘what if’ perspective, and not necessarily about optimization.
• Little’s law can be applied to single channel and multi channel queuing systems.
• When it comes arrival rates, there are two rates to account for: The arrival rate of customers asking for service, and the service delivery. If both rates are equal, then they are in balance. - 14What determines performance in queuing modelling?The availability of resources determines performance. If the rate of arrival is higher than the delivery of service, then waiting lines will be ever increasing. If the rate of arrivals is lower than the delivery of service, this means an under utilization of the servers.
- 15Five types of variability induced by customers - two are directly relevant to waiting lines1. Sources of variability: Arrival; customers do not want service at convenient for firm
2. Sources of variability: Request; service is an individual experience and there is a degree of variety in the delivery of it. - 16How to control arrival rates• It is important to consider arrival rates as an important variable; the number of arrivals over the hours that the service system is open for. Arrivals can be controlled using:
• Short lines
• Specific hours for specific customers
• “Specials”
--> The service time can be adapted by using faster or slower servers - 17Arrivals are sometimes assumed to arrive steadily, so that there is a constant arrival distribution and the exactly same time between successive arrivals periodically. What happens in case of a variable arrival population?• Arrivals are sometimes assumed to arrive steadily, so that there is a constant arrival distribution and the exactly same time between successive arrivals periodically.
In case of a variable arrival population:
1. Finite population: A limited-size customer pool that will use the service. At times they form a line. When a customer leaves its position as a member of the population, the size of the user group is reduced by one.
2. Infinite population: A population large enough that subtractions or additions to the population do not significantly affect system probabilities, such as banks, blood banks or pizza delivery. - 18Arrival patterns1. Exponential distribution: When arrivals at the service facility occur in purely random fashion (most likely for an infinite population).
2. Poisson distribution, or count distribution: Uses discrete values for an infinite population that is different than a continuous distribution. - 19Service time as constant or variable• A constant delivery of service is provided by automation, or more or less following pre-set scripts.
• A variable delivery of service is provided by humans and described using exponential distributions and other distributions. - 20What do waiting line models include?Waiting line models concerns the mathematical modelling of waiting lines, including several features: Structure (layout), phases (how many subsequent processes for delivery there are), statistical descriptions of arrival patterns, (statistical) descriptions of the delivery of service, and limitation to the lengths of the queue. There are four possible mathematical models in Jacobs and Chase (2014)
- 21Waiting line model 1: Simple system
- 22Waiting line model 2: Constant service time
- 23Waiting line model 3: Multiple channels
- 24Waiting line model 4: Finite customer system
- 25Probabilities for model 1• The probability of the length of queue being n is calculated using the Poisson distribution of arrivals and the exponential distribution of service time.
In the example on the slides, there is a 95% confidence that there are no more than 3 customers in the system. Thus, the business can either accept longer queues, improve processing or increase the number of servers.
• Waiting lines can be approximated. This method is alternative to calculating for Model 3. Approximation is based on observed data: The average and standard deviation of the interval between arrivals and the average and standard deviation of time for delivery of the service. Calculations are done in two steps: The calculation of variables and the calculation of the length of the queue and waiting line. - 26The complexity of waiting line problemsFor example, some have assumed waiting lines to be independent. When services become the input to the next process, simple formulas should no longer be used. This is also true for any problem where conditions do not meet the conditions for the equations. In this case, a computer simulation can be used.
- 27How customers can induce variability
- 28The effects of variabilityVariability leads to variations in demand in peak hours versus idle time, and due to queuing.
- 29Classic strategies for managing demandSome classic strategies for managing demand are the case strategy (responding to time and costs) and level strategy (smoothen demand)
Creating demand means to smoothen demand through offers or penalties. However, this strategy may have only limited effects. - 30Implications of queuing on capacity planning in service managementQueuing can have implications for capacity planning in service management, pertaining to the availability of resources based on service level (i.e. the percentage of customers served within a specific time period). Preferably, there should be short waiting / queuing times.
- 31How to mitigate effects of long queues• Concealing queues from arriving customers
• Using customers as a resource
• Making the customer’s wait comfortable and distracting attention
• Explain the reasons for the wait
• Provide pessimistic estimates of remaining waiting times
• Be open about queuing disciplines used
• Responding to arrival patterns:
• Flexibility in capacity utilization
• Feedforward as a control mechanism. For example, queuing in a supermarket and the number of checkout tills: Each queue hold maximum three customers and if longer, they get a shopping cart for free. This is an example on how to respond to fluctuations. NOTE: Nursing effect (whatever that is). - 32Summary: Key variables of queuing1. Competitive strategy: Waiting times and lengths of queues, where the number of servers is either identical or specialized.
2. Profitability: Return on investment on additional servers, where cost is indirectly related to competitiveness
Informações
- Descrição
- Business and Management 2A (Operations Management) FlashCards sobre 8. Queues and Queuing Systems, criado por Nina Olsson em 23-07-2018.
- Criado por
- Nina Olsson
- Visualizações
- 1
- Avaliação