Cloud computing is now widely used in all major industries. This transformative technology allows users to access the resources they need through a pay-as-you-go model, providing an efficient and convenient service. Although cloud computing can provide users with flexible resource scheduling and payment models, improper use may lead to cost overruns. In addition to makespan, the total cost of cloud services is also a key user requirement. Therefore, in this paper, we propose a heuristic algorithm aimed at solving the optimization problem of workflow scheduling with deadline constraints in cloud environments in order to minimize the total cost. The PACM algorithm assigns the corresponding deadline to each task in the workflow in upward probabilistic order, and achieves a more reasonable deadline assignment for the tasks based on the Critical Path Factor and Dependency Amplification Factor. Finally, the tasks are then assigned to cloud services to meet their sub-deadline requirements. Experiments are conducted using well-known scientific workflows for performance evaluation, and the results show that the algorithm outperforms previous heuristic algorithms, proving the effectiveness of the algorithm.

In heterogeneous embedded systems, we try to find a new non real-time scheduling algorithm, which can not only meet the reliability requirements, but also achieve low energy consumption. “Resource reservation time deterministic cyclic scheduling (RRTDCS)” algorithm is a kind of scheduling for offline cache pre-allocation. This scheduling algorithm combines time reservation and priority strategy. The algorithm can dynamically adjust according to different task sets. It can also adjust its own scheduling process to adapt to different scheduling environments. The periodic cyclic strategy enables low power consumption in task allocation. In the experimental part, we compared the algorithm proposed in this paper with the current mainstream real-time scheduling scheme. Experimental results show that RRTDCS algorithm has better performance.

Nuclei segmentation occupies an important position in medical image analysis, which helps to predict and diagnose diseases. With the further research of deep learning, the task of nuclei segmentation has been automated. However, most existing methods require a great deal of manually marked full masks for training, which is time-consuming and labor-intensive, and can only be done by professional personnel. For the purpose of reducing the cost of labeling, we propose a weakly supervised method using generative adversarial training for segmentation of nucleus. In the case of no boundary, but only the centroid of the nucleus, the proposed method segmented the nucleus region with blurred boundaries. We first use the generative adversarial network(GAN) to generate the likelihood map of the nuclear centroid, then use Guided Backpropagation to visualize the pixels that contributes to the detection of the centroid of each nucleus, and finally obtain the segmentation mask of the nucleus by graph-cut. In addition, for the purpose of training the network better, we performed stain normalization on each pathological image. We have verified the proposed method on a multi-organ nuclei dataset. The final experiment results show that our advanced method achieves better segmentation performance than other weakly supervised methods, and can even reach the level of full supervision.