M.V. Yashina,A.G. Tatashev
Abstract:A multi-channel queueing system is considered. The arriving requests differ in their type. Requests of each type arrive according to a Poisson process. The number of channels required for service with the rate equal to 1 depends of the request type. If a request is serviced with the rate equal to 1, then, by definition, the length of the request equals to the total service time. If at arrival moment, the idle channels is sufficient, then the arriving request is serviced with the rate 1. If, at the arrival moment, there are no idle channel, then the arriving request is lost. If, at arrival moment, there are idle channels but the number of idle channels is not sufficient for servicing with rate 1, then the request begins to be in service with rate equal to the ratio of the number of idle channels to the number of the channels required for service with the rate 1. If a request is serviced with a rate less than 1 and another request leaves the system, then the service rate increases for the request in consideration. Approximate formula for loss probability has been proposed. The accuracy of approximation is estimated. Approximate values are compared with exact values found from the system of equations for the related Markov chain stationary state probabilities.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the calculation of the request loss probability in the multi - channel queuing system. Specifically, the research object is a multi - channel queuing system, in which different types of requests arrive, and each type of request needs to occupy a certain number of channels for service. If there are not enough idle channels when the request arrives, the request will be rejected (i.e., lost). Therefore, accurately calculating the probability of request loss is crucial for understanding the performance of the system.
### Main problem description
1. **Multi - channel queuing system**:
- There are multiple channels in the system, and different types of requests arrive according to the Poisson process.
- Each type of request needs to occupy a certain number of channels for service, and the service rate can be 1 or less than 1.
- If there are enough idle channels when the request arrives, the request is served immediately; otherwise, the request may start to be served at a reduced service rate or be directly lost.
2. **Capacity sharing mechanism**:
- When a request arrives, if the number of idle channels is not enough to meet all the channels required by the request, the request will be served with partial channels, and the service rate will be reduced accordingly.
- If other requests are completed and release channels during the service process, the service rate of the current request will increase.
3. **Approximate calculation of loss probability**:
- Due to the complexity and high - dimensionality of the system state space, it is very difficult to accurately calculate the loss probability in practical applications.
- Therefore, the paper proposes an approximate method to estimate the loss probability and evaluates the accuracy of this approximate method.
### Main contributions of the paper
- Proposed an approximate formula to calculate the request loss probability in the multi - channel queuing system.
- Verified the effectiveness of the approximate method by comparing it with the exact values obtained from the relevant Markov chain steady - state probability equations.
- Analyzed the influence of different types of requests and service time distributions on the approximate results.
### Formula summary
The following key formulas are used in the paper:
1. **Inbound load**:
\[
A_i=\lambda_i b_i
\]
where \(\lambda_i\) is the arrival rate of the \(i\) - th type of request, and \(b_i\) is the average service time of the \(i\) - th type of request.
2. **Total load**:
\[
A = \sum_{i = 1}^{N}A_i
\]
3. **Weighted average number of required channels**:
\[
d=\frac{\sum_{i = 1}^{N}A_i d_i}{\sum_{i = 1}^{N}A_i}
\]
where \(d_i\) is the number of channels required for the \(i\) - th type of request.
4. **Equivalent number of channels**:
\[
v=\frac{n}{d}
\]
5. **Erlang formula**:
\[
c = E(A, v)=\frac{A^v e^{-A}}{\int_A^\infty y^{v - 1}e^{-y}dy}
\]
These formulas are used to calculate and evaluate the approximate loss probability and compare it with the exact solution to verify the accuracy of the approximate method.
### Conclusion
The paper proposes an approximate method to calculate the request loss probability in the multi - channel queuing system and verifies the effectiveness of this method through experiments. This is of great significance for understanding and optimizing the performance of multi - channel queuing systems in fields such as information networks and mobile communication systems.