High-level Scheduling Research Articles

Reservation table scheduling: branch-and-bound based optimizationvs. integer linear programming techniques

The recourse to operation research solutions has strongly increased the performances of scheduling task in the High-Level Synthesis (called hardware compilation). Scheduling a whole program is not possible as too many constraints and objectives interact. We decompose high-level scheduling in three steps. Step 1: Coarse-grain scheduling tries to exploit parallelism and locality of the whole program (in particular in loops, possibly imperfectly nested) with a rough view of the target architecture. This produces a sequence of logical steps, each of which contains a pool of macro-tasks. Step 2: Micro-scheduling maps and schedules each macro-task independently taking into account all peculiarities of the target architecture. This produces a reservation table for each macro-task. Step 3: Fine-grain scheduling refines each logical step by scheduling all its macro-tasks. This paper focuses on the third step. As tasks are modeled as reservation tables, we can express resource constraints using dis-equations (i.e. , negations of equations). As most scheduling problems, scheduling tasks with reservation tables to minimize the total duration is NP-complete. Our goal here is to design different strategies and to evaluate them, on practical examples, to see if it is possible to find optimal solution in reasonable time. The first algorithm is based on integer linear programming techniques for scheduling, which we adapt to our specific problem. Our main algorithmic contribution is an exact branch-and-bound algorithm, where each evaluation is accelerated by variant of Dijkstra's algorithm. A simple greedy heuristic is also proposed for comparisons. The evaluation and comparison are done on pieces of scientific applications from the PerfectClub and the HLSynth95 benchmarks. The results demonstrate the suitability of these solutions for high-level synthesis scheduling.

Hybrid Preemptive Scheduling of Message Passing Interface Applications on Grids

Time sharing between cluster resources in a grid is a major issue in cluster and grid integration. Classical grid architecture involves a higher-level scheduler which submits non-overlapping jobs to the independent batch schedulers of each cluster of the grid. The sequentiality induced by this approach does not fit with the expected number of users and job heterogeneity of grids. Time sharing techniques address this issue by allowing simultaneous executions of many applications on the same resources. Co-scheduling and gang scheduling are the two best known techniques for time sharing cluster resources. Coscheduling relies on the operating system of each node to schedule the processes of every application. Gang scheduling ensures that the same application is scheduled on all nodes simultaneously. Previous work has proven that coscheduling techniques outperform gang scheduling when physical memory is not exhausted. In this paper, we introduce a new hybrid sharing technique providing checkpoint-based explicit memory management. It consists in co-scheduling parallel applications within a set, until the memory capacity of the node is reached, and using gang scheduling related techniques to switch from one set to another one. We compare experimentally the merits of the three solutions (co-scheduling, gang scheduling, and hybrid scheduling) in the context of out-of-core computing, which is likely to occur in the grid context, where many users share the same resources. Additionally, we address the problem of heterogeneous applications by comparing hybrid scheduling to an optimized version relying on paired scheduling. The experiments show that the hybrid solution is as efficient as the co-scheduling technique when the physical memory is not exhausted, can benefit from the paired scheduling optimization technique when applications are heterogeneous, and is more efficient than gang scheduling and coscheduling when physical memory is exhausted.

Joint application mapping/interconnect synthesis techniques for embedded chip-scale multiprocessors

As transistor sizes shrink, interconnects represent an increasing bottleneck for chip designers. Several groups are developing new interconnection methods and system architectures to cope with this trend. New architectures require new methods for high-level application mapping and hardware/software codesign. We present high-level scheduling and interconnect topology synthesis techniques for embedded multiprocessor systems-on-chip that are streamlined for one or more digital signal processing applications. That is, we seek to synthesize an application-specific interconnect topology. We show that flexible interconnect topologies utilizing low-hop communication between processors offer advantages for reduced power and latency. We show that existing multiprocessor scheduling algorithms can deadlock if the topology graph is not strongly connected, or if a constraint is imposed on the maximum number of hops allowed for communication. We detail an efficient algorithm that can be used in conjunction with existing scheduling algorithms for avoiding this deadlock. We show that it is advantageous to perform application scheduling and interconnect synthesis jointly, and present a probabilistic scheduling/interconnect algorithm that utilizes graph isomorphism to pare the design space.

Hierarchical Dynamic Bandwidth Allocation Algorithm for Multimedia Services over Ethernet PONs

In this paper, we propose a new dynamic bandwidth allocation (DBA) algorithm for multimedia services over Ethernet PONs (passive optical networks). The proposed algorithm is composed of a low-level scheduler in the optical network unit (ONU) and a high-level scheduler in the optical line terminal (OLT). The hierarchical DBA algorithm can provide expansibility and efficient resource allocation in an Ethernet PON system in which the packet scheduler is separated from the queues. In the proposed DBA algorithm, the OLT allocates bandwidth to the ONUs in proportion to the weight associated with their class and queue length, while the ONU preferentially allocates its bandwidth to queues with a static priority order. The proposed algorithm provides an efficient resource utilization by reducing the unused remaining bandwidth caused by the variable length of the packets. We also define the service classes and present control message formats conforming to the multi-point control protocol (MPCP) over an Ethernet PON. In order to evaluate the performance, we designed an Ethernet PON system on the basis of IEEE 802.3ah “Ethernet in the first mile” (EFM) using OPNET and carried out simulations. The results are analyzed in terms of the channel utilization, queuing delay, and ratio of the unused remaining bandwidth.

Resource-Constrained Shop-Level Scheduling in a Shipyard

Ship production planning and scheduling at higher levels do not explicitly consider scheduling details at the level of individual workshops. However, the schedule of major events in ship production is collectively influenced by the actual shop-level, short-interval production schedules, which depend on resource and material availability and also on the due dates and priorities of the workloads. This necessitates development of robust, resource-constrained, shop-level scheduling systems that can support higher-level schedules in ship production. WorkShip (Knowledge Based Systems, Inc., College Station, TX) is a software tool for scheduling workload over short, regular intervals in workshops of a shipyard. It is driven by a powerful scheduling engine that is based on a generic model of resource-constrained job-shop scheduling and an efficient scheduling technique. Similar scheduling systems are being developed in other shops so that all systems can be used in tandem to support higher-level scheduling and help achieve optimal productivity for the shipyard.

Using Mobile Agents to Schedule a Manufacturing Chain on the Internet

This paper is mainly concerned with putting forward a novel manufacturing chain model and a new scheduling model for manufacturing chains in networked manufacturing systems. The manufacturing chain model is formed by adopting extensive activity-network-diagram workflow technique. The scheduling model is built upon mobile agent technologies, e.g. aglets. It includes two different scheduling levels. One is defined as the process scheduling which implements the scheduling logic for distributive manufacturing sites or manufacturing cells based on a specific manufacturing task. The other is the node scheduling which realizes the scheduling logic for manufacturing equipments based on specific working procedures or working paces. High-level scheduling aglet templates are created to support the operations at the levels of manufacturing sites or manufacturing cells. To complement, low-level scheduling aglet templates are created to support the operations such as working procedures or working paces over manufacturing equipments. As a result, the model is able to deal with dynamic manufacturing chains and their scheduling.

Convergent matching for model-based computer vision

We consider matching in model-based computer vision as a converging discrete iteration and give a basis for examining the convergence as the movement of the working point in a lattice. Because the matching is non-deterministic we discuss convergence in terms of completing sub-problems within a time slot. This form of low-level scheduling avoids effectively unlimited trials of sub-graphs, a phenomenon that we call the NP-trap. We define high-level scheduling as the need to test each reference class at least once and thereafter focus attention on the most promising candidates. Examples show the bounding of matching time with a time slot and focusing of attention guided by a figure of merit.

EDAR – mobile robot for parts moving based on a game-theoretic approach

EDAR (event-driven assembler robot) – a mobile robot capable of moving a collection of disk-shaped parts located on a two-dimensional workspace from an arbitrary initial configuration to a desired configuration while avoiding collisions in a purely reactive manner, is presented. Since EDAR uses a higher-level scheduler to switch among the subtasks of moving individual parts, it is viewed as mediating a noncooperative game played among the parts.

Experimental investigation of iterative simulation-based scheduling in a dynamic and stochastic job shop

Abstract A vital component of modern manufacturing systems is the scheduling and control system, which determines companies' overall performance in their respective supply chains. This paper studies iterative simulation-based scheduling mechanisms for manufacturing systems that operate in dynamic and stochastic environments. Also assessed are the issues involved when these mechanisms are used to make higher-level scheduling decisions, such as dispatching rule selection, instead of generation of a full schedule. A typical simulation-based system is outlined and tested under various experimental conditions. Examined are the effects of stochastic events such as machine breakdowns and processing time variations on the system performance, and the effectiveness of the simulation-based approach from the control point of view is evaluated. Finally, different levels of two important factors (look-ahead window and scheduling period) are compared for the iterative approach. Computational results show that, although simulation-based scheduling proves effective when these parameters are properly set, the overall performance diminishes due to the dynamic and stochastic nature of the system, which degrades the multi-pass improvement capability of the simulation runs. Experimental results also support the initial expectation in that frequent updates to the higher-level schedule may not be necessary when these decisions are naturally “adaptive” to the unexpected system changes.

Efficient module selections for finding highly acceptable designs based on inclusion scheduling

Abstract In high level synthesis, module selection, scheduling, and resource binding are inter-dependent tasks. For a selected module set, the best schedule/binding should be generated in order to accurately assess the quality of a module selection. Exhaustively enumerating all module selections and constructing a schedule and binding for each one of them can be extremely expensive. In this paper, we present an iterative framework, called W i Z ard to solve module selection problem under resource, latency, and power constraints. The framework associates a utility measure with each module. This measurement reflects the usefulness of the module for a given a design goal. Using modules with high utility values should result in superior designs. We propose a heuristic which iteratively perturbs module utility values until they lead to good module selections. Our experiments show that by keeping modules with high utility values, W i Z ard can drastically reduce the module exploration space (approximately 99.2% reduction). Furthermore, the module selections formed by these modules belong to superior solutions in the enumerated set (top 15%).

A Hybrid Procedural/Deductive Executive for Autonomous Spacecraft

The New Millennium Remote Agent (NMRA) will be the first AI system to control an actual spacecraft. The spacecraft domain places a strong premium on autonomy and requires dynamic recoveries and robust concurrent execution, all in the presence of tight real-time deadlines, changing goals, scarce resource constraints, and a wide variety of possible failures. To achieve this level of execution robustness, we have integrated a procedural executive based on generic procedures with a deductive model-based executive. A procedural executive provides sophisticated control constructs such as loops, parallel activity, locks, and synchronization which are used for robust schedule execution, hierarchical task decomposition, and routine configuration management. A deductive executive provides algorithms for sophisticated state inference and optimal failure recovery planning. The integrated executive enables designers to code knowledge via a combination of procedures and declarative models, yielding a rich modeling capability suitable to the challenges of real spacecraft control. The interface between the two executives ensures both that recovery sequences are smoothly merged into high-level schedule execution and that a high degree of reactivity is retained to effectively handle additional failures during recovery.

Large Scale System Issues in Automated Highway and Air Traffic Management Systems

Large scale systems typically consist of compositions of many interacting sub-systems performing various planning and control functions. Control for such systems is often arranged in a hierarchical structure, with high level scheduling functions concerned with optimality and efficiency issues and low level control functions implementing the desired high level objectives and dealing with system uncertainty and safety. This paradigm leads to interactions between the fields of optimization and control theory. In this paper, these issues are highlighted using two large scale applications: Automated Highway Systems (AHS) and Air Traffic Management (ATM) Systems.

A two-level hierarchical approach for raw material scheduling in steelworks

Raw material scheduling problems in steelworks suffer from the computational complexity arising from the interdependence and interference between various items of transport equipment. To deal effectively with such complexity, this paper proposes a two-level hierarchical architecture for the construction of predictive schedules to transport raw materials from stock yards to processing plants. The architecture allocates the raw materials required, to multiple lower-level dispatchers, each of which schedules its reclaimer to withdraw the assigned raw materials. A higher-level scheduler constructs belt-conveyor schedules for transporting the reclaimed raw materials to their processing plants. If any conflict occurs within a dispatcher during the scheduling process, the higher-level scheduler coordinates other dispatchers for conflict resolution. Based on this architecture, a scheduling expert system called RAMS was developed using G2, to automate the raw material scheduling process at a steelworks. By providing the predictive raw material transport schedule, the system is proved to relieve the burden on the human scheduler and to improve schedule performances.

Knowledge based tools for hubble space telescope planning and scheduling: Constraints and strategies

The Hubble Space Telescope (HST) presents an especially challenging scheduling problem since a year's observing program encompasses tens of thousands of exposures facing numerous coupled constraints. This article discusses recent progress in the development of planning and scheduling tools which augment the existing HST ground system. General methods for representing activities, constraints and constraint satisfaction, and time segmentation have been implemented in a scheduling testbed. The testbed permits planners to evaluate optimal scheduling time intervals, calculate resource usage and to generate long- and medium-range plans. Graphical displays of activities, constraints and plans are an important feature of the system. High-level scheduling strategies using rule based and neural net approaches have been implemented.