#queueing-theory

At high utilization levels, losing capacity generates a queue much faster than regaining capacity can shrink.

Each time you halve the amount of excess capacity, you double the queue. You fold the queue as you move from 80 to 90 or 90 to 95 percent.

The percent of the time that the arriving work will find the server busy, the average number of items in the queue, the average number of items in the system.

Queues in high-state are the same as congested queues, queues at high capacity utilization, with high WIP, or at high level of utilization

To make supply-focused approaches effective, you must minimize the time between the onset of congestion and the application of added resources.

One of the biggest advantages of the demand-focused approaches is the speed at which you can execute them.

When WIP becomes high, you can purge WIP by decreasing the batch size of your jobs. A way to achieve this is to shed the requirements of your jobs.

When WIP is high, purge low-value projects from the queue to make room for high-value projects.

Whenever WIP is high, queues are high. When queues are high, jobs in the process generate high delay costs. Most of the damage is caused by high-state queues.

You can block queues in two ways: 1. Eject the arriving job from the system. 2. Hold back the job in an upstream queue.

Queues create many forms of economic waste: longer cycle time, increased risk, more variability, more overhead, lower quality, and less motivation.

Low-queue states are more probable than high-queue states, but high-queue states are more important because they delay more jobs.

So when your system is operating at maximum capacity utilization, you don't have slack. When it isn't working at its full, you have slack.

If the time on the critical path is worth a lot of money and a support group is on the path, you should measure the group on response time, not efficiency.

In this situation, if your team can execute the tasks of the bottleneck process, they can help it by increasing its capacity for a short period.

Because of the pull principle, the upstream processes can't send work to the next stage if a blockage occurs anywhere.

When you set a limit on WIP, you create positive and negative effects.

A random process (sequence of random variables) has the chance to spin out from the mean because the variance of the accumulative total grows larger with time.

The idle time of waiting works in queues increases the work-in-process, which is the root cause of much waste in product development.

A high number of working in the process leads to a high queue capacity utilization, which causes problems such as a higher cycle time.

With this algorithm, the system handles low-priority and doesn't get stuck in large jobs with uncertain time to complete.

Cost of delay (COD) determines the financial impact on the company if it releases a product on the market earlier or later. When a product arrives on the market X months earlier, revenues also roll in X months earlier.

The law states that average queue time will equal the average queue size divided by the average processing rate.

The cumulative flow diagram (CFD) is a helpful tool for managing queues.

In queueing theory, an M/G/1 queue is a queue in which the arrivals are a Markovian, service has a general distribution, and a single server.

Queue capacity utilization is the amount of time the service process is working (not idle).

The queue service process is the process in which the server accomplishes the work.

The critical path is the most extended sequence of activities in a project plan. The project is only completed on time only if you meet these activities.

It can provide essential insights to product developers because there're similar problems of unpredictable work arrival time and unpredictable task durations.