91Ó°ÊÓ

In queueing theory, a common model is the M/G/1 queue. This model represents a single-server queueing system.
Here, 'M' stands for Markovian (or memoryless) arrival process, 'G' denotes a general service time distribution, and '1' indicates a single server. In the M/G/1 model, customers arrive at the queue according to a Poisson process.
This means the time between arrivals is exponentially distributed. However, the service times can take on any distribution.
The flexibility of the service time distribution allows for a more realistic description of many real-world systems.

• Important: Arrival process is Poisson.
• Service time can be any distribution.
• Single-server handles one customer at a time.

Understanding this model helps in analyzing and predicting system performance, such as waiting time and queue length.

The utilization factor, denoted by \(\rho\), is a crucial metric in the study of queueing systems. It represents the proportion of time the server is actively working.
Mathematically, it is defined as the ratio of the arrival rate \(\lambda\) to the service rate \(\mu\): \(\rho = \frac{\lambda}{\mu}\).

• \(\lambda\): Arrival rate (average number of arrivals per unit time).
• \(\mu\): Service rate (average number of customers served per unit time).
• \(\rho < 1\): The system is stable.

This factor indicates how busy the server is. If \(\rho\) equals 1 or more, the system may become overloaded, leading to infinite queues.
Therefore, maintaining a utilization factor less than 1 is essential for a stable system.

The Pollaczek-Khintchine (P-K) formula is a fundamental tool in queueing theory.
It provides a way to calculate the average number of customers in the system, denoted by \(L\), in an M/G/1 queue.
The formula is given by:\[ L = \rho + \frac{\lambda^2 \sigma^2 + \lambda^2 \mu^2}{2(1-\rho)} \]

• \(\rho\): Utilization factor.
• \(\lambda\): Arrival rate.
• \(\sigma^2\): Variance of the service time.
• \(\mu\): Service rate.

This formula shows that the average number of customers in the system depends on both the variability of the arrival and service processes.
It highlights the importance of service time variability (\(\sigma^2\)) in influencing queue length.
A higher variance indicates more fluctuations, leading to longer queues and wait times.

An idle period in a queueing system is the time when the server is not busy and there are no customers in the queue.
This concept helps in understanding how often a server is completely free from work. In the M/G/1 queue, the probability that the system is idle is:\[ P_0 = 1 - \rho \]

• This means idle period probability is inversely related to \(\rho\).
• More idle periods occur when \(\rho\) is small.
• During idle periods, the server waits for new arrivals.

This probability, \(P_0\), helps to determine the proportion of departures that leave behind zero work.
A high portion of idle time means that many departures occur without leaving customers behind.

The concept of average work in a system is essential for understanding performance metrics in queueing theory.
In an M/G/1 queue, it refers to the average amount of work or time a customer spends in the system.The average work can be calculated by dividing the average number of customers in the system (\(L\)) by the arrival rate (\(\lambda\)):\[ W = \frac{L}{\lambda} \]

• \(W\): Average time a customer spends in the system.
• \(L\): Average number of customers in the system (from P-K formula).
• \(\lambda\): Arrival rate.

This measure provides insight into system efficiency.
A lower average work time signifies a more efficient system with quicker service and shorter queues.