Abstract:The rapid development of the mobile Internet and the Internet of Things is leading to a diversification of user devices and the emergence of new mobile applications on a regular basis. Such applications include those that are computationally intensive, such as pattern recognition, interactive gaming, virtual reality, and augmented reality. However, the computing and energy resources available on the user's equipment are limited, which presents a challenge in effectively supporting such demanding applications. In this work, we propose a heterogeneous computing resource allocation model based on a data-driven approach. The model first collects and analyzes historical workload data at scale, extracts key features, and builds a detailed data set. Then, a data-driven deep neural network is used to predict future resource requirements. Based on the prediction results, the model adopts a dynamic adjustment and optimization resource allocation strategy. This strategy not only fully considers the characteristics of different computing resources, but also accurately matches the requirements of various tasks, and realizes dynamic and flexible resource allocation, thereby greatly improving the overall performance and resource utilization of the system. Experimental results show that the proposed method is significantly better than the traditional resource allocation method in a variety of scenarios, demonstrating its excellent accuracy and adaptability.
What problem does this paper attempt to address?
The problems that this paper attempts to solve are as follows: Against the backdrop of the rapid development of the mobile Internet and the Internet of Things (IoT), user devices are becoming more diverse and new mobile applications keep emerging. In particular, computationally - intensive applications (such as pattern recognition, interactive games, virtual reality, and augmented reality) have an ever - increasing demand for computing resources and energy. However, the computing and energy resources of user devices are limited, making it difficult to effectively support these high - demand applications. For this reason, the paper proposes a heterogeneous computing resource allocation model based on a data - driven method.
Specifically, this paper aims to solve the following problems:
1. **Limited computing and energy resources**: The hardware limitations of user devices lead to limited computing resources, storage resources, and energy resources, making it difficult to efficiently support computationally - intensive and low - latency - sensitive applications.
2. **Limitations of cloud computing**: Although computing tasks can be offloaded to the cloud for processing, a large amount of data transmission and the distance between cloud servers and users will lead to high - transmission latency and reliability problems, affecting the user experience.
3. **Challenges of edge computing**: With the development of mobile edge computing, different types of edge servers need to work together to process computing tasks, which involves computing task offloading, multi - dimensional resource allocation, and sharing of computing results. At the same time, security issues of malicious nodes that may exist in the network also need to be addressed.
4. **Optimization of heterogeneous computing systems**: In order to improve energy efficiency and computing performance, more and more computing systems adopt heterogeneous architectures (such as CPU, GPU, DSP, FPGA, etc.). How to achieve efficient resource allocation in such a heterogeneous environment is an important topic.
To solve these problems, the paper proposes a data - driven deep neural network prediction model. By collecting and analyzing historical workload data, extracting key features and constructing a detailed data set, it predicts future resource requirements and dynamically adjusts and optimizes the resource allocation strategy according to the prediction results. This model not only fully considers the characteristics of different computing resources but also accurately matches the requirements of various tasks, achieving dynamic and flexible resource allocation, thereby significantly improving the overall performance and resource utilization of the system.
### Key Formulas
1. **Moving average derivative feature**:
\[
x_{t,N + 1}=\frac{1}{k}\sum_{i = 0}^{k - 1}CPU_{t - i}
\]
where \(k\) is the size of the moving window.
2. **Mean - square - error loss function**:
\[
L=\frac{1}{T}\sum_{t = 1}^{T}(r_{t + 1}-\hat{r}_{t + 1})^2
\]
where \(T\) is the number of samples, \(r_{t+1}\) is the actual resource requirement, and \(\hat{r}_{t + 1}\) is the predicted resource requirement.
3. **System utility function**:
\[
U=\sum_{i = 1}^{N}(\alpha_iTET_i+\beta_iEnergy_i)
\]
where \(N\) is the number of tasks, \(\alpha_i\) and \(\beta_i\) are the time weight and energy consumption weight of the \(i\) - th task respectively, \(TET_i\) is the task execution time, and \(Energy_i\) is the user energy consumption.
4. **Joint optimization problem**:
\[
\min_{u,p,b}\sum_{i = 1}^{N}\left(\alpha_i\left(\sum_{j = 1}^{M}u_{ij}TET_{ij}+(1-\sum_{j = 1}^{M}u_{ij})TET_{i,local}\right)+\beta_i\left(\sum_{j = 1}^{M}u_{ij}Energy_{ij}+(1-\sum_{j = 1}^{M}u_{ij})Energy_{i,local}\right)\right)
\]
where \(M\) is the number of edge servers, \(u_{ij}\) represents the...