7results about How to "Improve schedulability" patented technology

An electric vehicle collaborative power grid power supply recovery strategy considering multi-time-domain EV-EPSV comprises the steps that a multi-time-sequence electric vehicle EV equivalent power supply model is established according to a travel rule of an electric vehicle user in a typical travel scene; a load importance index system is constructed, a load power loss model is established in combination with the load capacity and the power failure duration, and quantitative calculation of economic and social losses of different types of loads in the multi-time-sequence power failure process is achieved; and constructing a collaborative optimized power supply model of the electric vehicle EV and the emergency power supply vehicle EPSV, and realizing static power distribution of the electric vehicle and dynamic path and power optimization scheduling of the emergency power supply vehicle. According to the strategy, a multi-time-sequence EV equivalent power supply model is established, a load power loss evaluation system is constructed, and a collaborative optimization scheduling method of the electric vehicle and the emergency power supply vehicle is provided, so that efficient distribution of limited power supply resources and improvement of power supply recovery efficiency are realized. The scheduling flexibility of electric vehicle resources and the power supply toughness of a power grid can be effectively improved in a sudden power failure situation.

This invention discloses an online scalable configuration method and system for dynamically changing time-sensitive networks (TSNs), relating to the Internet of Things (IoT) field. The invention proposes an efficient inter-flow conflict detection mechanism based on multi-level flow grouping. By using a correlation analysis of flow period and offset, it avoids traditional conflict detection methods based on link maximal cliques and link hyperperiods, enabling rapid response to large-scale data flow conflict detection needs. It designs an online incremental method based on decoupling routing and scheduling, as well as offline pre-routing, accelerating the computation speed of routing and scheduling schemes while ensuring scheduling space. This invention reduces the computational complexity of data flow set hyperperiod sensitivity and the maximum link slot occupancy in TSNs, accelerating online scheduling and providing efficient conflict detection support for online scalable configuration.

The application discloses a method for layer scheduling of mixed deep neural network tasks in an embedded real-time system. The method considers the limited CPU and GPU resources of the embedded real-time system, the basic conditions of task scheduling, the time limit of task scheduling and other factors, and designs a scheduling mechanism for layer distribution of mixed deep neural network tasks for the embedded real-time system. The method builds a deep neural network task model and a layer task model, mathematically calculates the task scheduling overhead, the worst response time, the utilization sum of the task under the minimum layer task mapping scheduling, sets an optimization function and a constraint condition solving method, and realizes more balanced utilization of the heterogeneous CPU and GPU computing resources in the real-time system under the condition of shorter worst response time, and improves the real-time performance and schedulability of deep neural network tasks in the embedded real-time system.

The application discloses a kind of cross TSN and PON domain deterministic cooperative routing and scheduling method, by TDM-PON is equivalent to the TSN switch (i.e. OTSN) with TAS function, the cooperative scheduling of OTSN and TSN equipment is realized, the E2E deterministic transmission problem is solved.And the TW constraint formula for OTSN and TSN joint scheduling and the cross-domain cooperative routing and scheduling scheme based on the greedy algorithm are proposed to plan the position of TW, meet the delay, jitter transmission demand of various businesses, compared with traditional non-cooperative scheduling scheme, the schedulability is improved, it is a kind of E2E deterministic transmission capacity, meet the routing and scheduling scheme of the demand of multiple types of businesses.

A multi-time-sequence EV equivalent power supply model modeling method considering a vehicle travel rule comprises the following steps: according to dynamic evolution characteristics of the travel rule of an electric vehicle user, combining travel sample data, dividing into a summer scene and a winter scene according to seasons, and dividing into a workday scene and a holiday scene according to a time sequence to complete data fitting; then, a vehicle travel state transition matrix is established through a semi-Markov chain SMC, and a Monte Carlo method MCM is utilized to perform random simulation on vehicle behaviors to obtain available discharge capacities of the electric vehicles at all moments, so that an electric vehicle cluster in a region is equivalent to a time-varying power supply model, and a multi-time-sequence EV equivalent power supply model is established; according to the method, by introducing a semi-Markov chain and a Monte Carlo random simulation method, available energy storage capacity distribution characteristics of an electric vehicle group in different scenes are described. The model can reflect energy storage laws of residential areas, working areas and shopping areas in different seasons and days, and dynamic quantitative evaluation of the EV available power potential is realized.

The invention relates to a multi-core real-time system task sorting and partition scheduling optimization method based on a DQN, and the method comprises the following steps: obtaining a real-time task set to be scheduled, and extracting the execution time, period, deadline and other feature information of tasks; modeling a task sorting process as a Markov decision process, and performing adaptive optimization on a task processing sequence by using a DQN model to generate a task sorting sequence; on the premise that the structure of a partition scheduling heuristic algorithm is not changed, the task sorting sequence is input into the partition scheduling algorithm, and tasks are distributed; and generating scheduling feedback information according to a task allocation result, wherein the scheduling feedback information is used for updating the DQN model. According to the method, intelligent optimization is carried out on the task sorting stage, the problem that a traditional fixed sorting strategy is difficult to adapt to a complex task set structure is solved, a partition scheduling algorithm can find a feasible task allocation scheme under the high system load condition, and therefore the schedulability and the scheduling success rate of the isomorphic multi-core real-time system are improved.

This invention relates to the field of artificial intelligence and discloses a method and system for scheduling computing resources in heterogeneous computing systems based on intelligent prediction. This solution obtains task description information from a queue of tasks to be scheduled, analyzes task resource requirements, topological constraints, and portability attributes, and classifies tasks into high-resource-demand tasks, ordinary tasks, and elastic tasks. Based on task category, resource fragmentation level, and the arrival status of high-resource-demand tasks, ordinary tasks and elastic tasks are allocated to fragmented partitions, while high-resource-demand tasks are allocated to reserved partitions. Continuous resource blocks matching the resource requirements of high-resource-demand tasks are reserved in the reserved partitions, generating scheduling decisions. Based on the scheduling decisions and the updated reserved and fragmented partitions, task start, pause, migration, and termination operations are executed, and operational feedback information is collected and written to an operational data repository. This improves the schedulability of high-resource-demand tasks.