Parallel Production Systems Research Articles

Optimal Scheduling in Parallel and Serial Manufacturing Systems via the Maximum Principle

The problems of M-machine, J-product, N-time point preemptive scheduling in parallel and serial production systems are the focus of this paper. The objective is to minimize the sum of the costs related to inventory level and production rate along a planning horizon. Although the problem is NP-hard, the application of the maximum principle reduces it into a well-tractable type of the two-point boundary value problem. As a result, algorithms of O(NMJk(N-L)+1) and O(N(MJ)k(N-L)+1) time complexities are developed for parallel and serial production systems, respectively, where L is the time point when the demand starts and k is the ratio of backlog cost MN over the inventory cost. This compares favorably with the time complexity O((J+1MN) of a naive enumeration algorithm.

An analytical formula for variance of output from a series-parallel production system with no interstation buffers and time-dependent failures

This paper presents a method to determine the mean and the variance of the amount of materials produced in a fixed time interval by a continuous materials flow production system with N stations in series and M stations in parallel and no interstation buffers. Unreliable stations with exponential failure and repair times, time dependent failures, and deterministic processing times are considered. Closed-form expressions for the asymptotic mean and variance of the amount of materials produced per unit time are given for series, parallel, and series-parallel production systems with identical stations. It is shown that the distribution of the amount of materials produced in a fixed time is asymptotically normal. By using this property, effects of variability on the due-date performance are investigated by considering the probability of meeting a customer order on time. Numerical experiments that explore some relationships among performance measures and production system parameters are also presented.

Exploiting application parallelism in knowledge-based systems: an experimental method

A method is presented which aims to enhance the run-time performance of real-time production systems of utilising natural concurrency in the application knowledge base. This exploiting application parallelism (EAP) method includes an automated analysis of the knowledge base and the use of this analysis information to partition and execute rules on a novel parallel production system (PPS) architecture. Prototype analysis tools and a PPS simulator have been developed for the Inference ART environment in order to apply the method to a naval data-fusion problem. The results of this experimental investigation revealed that an average maximum of 12.06 rule-firings/cycle was possible but, due to serial bottlenecks inherent in the data-fusion problem, up to only 2.14 rule-firings/cycle was achieved overall. Limitations of the EAP method are discussed within the context of the experimental results and an enhanced method is investigated.

An implementation of parallel production systems with two load balancing mechanisms

AbstractAn implementation method is proposed for parallel production systems on multicomputers, or message‐passing computers, to speed up execution time. There have been proposed parallel production systems using hash mechanism, but they cause a skewed load distribution problem. To obtain more efficient balance of load, the method addressed here and named clustered parallel production systems (CPPS) adopts two load balancing strategies: hash and demand‐driven. Taking account of the cost of termination detection, the execution time of the CPPS and simple hash method are estimated, and implement on an nCube2. The estimation meets the execution results. The CPPS provides much better load balance to improve scalability.

Issues in the Design and Control of Parallel Rule-Firing Production Systems

The parallel execution of rules in a production system provides the potential for faster execution, but increases the complexity of control and design issues. We address two issues: controlling the execution of productions without introducing serial bottlenecks and maintaining correctness during the course of simultaneous rule executions. Two novel rule-firing policies are described: an asynchronous rule-firing policy that causes rules to be executed as soon as they become enabled and a task-based scheduler that allows multiple independent tasks to run asynchronously with respect to each other while allowing rules to execute either synchronously or asynchronously within the context of each task. Previous research in parallel rule-firing systems has indicated that a serializable result cannot be guaranteed without a run-time mechanism for detecting potentially harmful rule interactions. Our analysis of such mechanisms indicates that their overhead is prohibitive for asynchronous rule-firing systems. In exchange for improved performance, we trade the guarantee of serializability for the somewhat weaker claim that correct parallel rule-firing programs may be designed, given the appropriate language mechanisms. We present a simple locking scheme for working memory, which, when coupled with the appropriate language idioms, allows serializable programs to be developed without incurring the expense of run-time interference detection. The experimental results of this research are presented in the context of UMass Parallel OPS5, a rule-based language that incorporates parallelism at the rule, action, and match levels, and provides language constructs for supporting the design of parallel rule-based programs. Results are presented for a number of programs illustrating common AI paradigms including search, inference, and constraint satisfaction problems.

A TRANSACTION MODEL FOR PARALLEL RULE EXECUTION PART I: MODEL AND ALGORITHMS

This paper introduces a transaction model for parallel production systems, where rules can be fired in a completely asynchronous parallel manner. A model is developed to describe two properties of a serial production system execution, namely, conditional dependence (a rule firing removes another from the conflict set) and prioritization (by the conflict resolution mechanism). These are used as the basis to define the correctness criteria for parallel production execution, namely, conditional serializability, prioritized serializability, and prioritized conditional serializability. A graph model is used to represent production system execution which provides a means of expressing correctness as graph acyclicity, which in turn is used as a basis to develop practical schedulers that manage the concurrent asynchronous production execution and ensure correctness. Two practical schedulers, namely, conditional strict two-phase locking and priority queue scheduler are developed for conditional serializability and prioritized serializability, respectively. They are combined for prioritized conditional serializability.

A TRANSACTION MODEL FOR PARALLEL PRODUCTION SYSTEMS PART II: MODEL AND EVALUATION

Parallel-rule firing approaches have been proposed to improve the performance of production systems. However, few models have been developed to measure the performance of parallel-rule firing approaches. There are three approaches to evaluating system performance, namely, analytical modeling, simulation, and system prototyping or implementation. Analytical modeling has its advantage in being cost effective and providing a powerful tool for studying various aspects of system performance by varying system parameters. Since analytical modeling has been very successfully used in evaluating database system performance, we develop such a model for parallel production systems, where rule firing is modeled as a transaction. Both resource contention and data contention are modeled in detail and the performance of locking, timestamp and optimistic approaches is analyzed. We show that significant speedup can be gained in parallel rule execution. Our main contribution is in the insights into parallel-rule firing provided by parametric modeling.

A SURVEY ON THE PARALLEL DISTRIBUTED PROCESSING OF PRODUCTION SYSTEMS

The importance of production systems in artificial intelligence (AI) has been repeatedly demonstrated by a large number of expert systems. As the number and size of expert systems grow, there has however been an emerging obstacle in such AI applications: the large processing time. The need for faster execution of production systems has spurred research in both the software and hardware domains, including connectionist architectures. This paper surveys various aspects of parallel distributed processing of production systems. Approaches taken to date to solve the problems associated with production systems are classified here into three levels: the algorithmic level, the parallel implementation level, and the connectionist level. Several pattern matchers and multiple rule firing principles are presented to demonstrate the algorithm level improvement. Several parallel implementation efforts are surveyed along with experimental results on real machines or with simulators. The presentation of three different types of connectionist production systems (local, distributed, and hierarchical representation) completes this survey. Finally, we explore some potential avenues towards the implementation of a true asynchronous parallel production system.

The state of the art in parallel production systems

The production system paradigm occupies a prominent place in artificial intelligence. Production systems have not yet been widely accepted in industry mainly due to their slow performance. Continuing research in knowledge processing requires larger and larger production systems, which would only exacerbate the performance problem. For this reason, it is important to apply parallel processing technology to production systems because it may provide the speed improvement necessary for future production systems. This paper examines recent research efforts in production systems. It begins by discussing the architecture of production systems and the cause of their slow performance. It groups the research efforts into three categories, faster sequential match algorithms, parallel match production systems, and multiple rule firing production systems, and analyzes the strength and weakness of each approach. A uniform terminology is used throughout the paper. By considering each category individually and comparing them collectively, a clear picture of recent research efforts in production systems is obtained.

Organization self-design of distributed production systems

The authors introduce two reorganization primitives, composition and decomposition, which change the population of agents and the distribution of knowledge in an organization. To create these primitives, they formalize organizational knowledge, which represents knowledge of potential and necessary interactions among agents in an organization. The authors develop computational organizational self-design (OSD) techniques for agents with architectures based on production systems to take advantage of the well-understood body of theory and practice. They first extend parallel production systems, where global control exists, into distributed production systems, where problems are solved by a society of agents using distributed control. Then they introduce OSD into distributed production systems to provide adaptive work allocation. Simulation results demonstrate the effectiveness of the approach in adapting to changing environmental demands. The approach affects production system design and improves the ability of build production systems that can adapt to changing real-time constraints.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A parallel production system architecture

The growing importance of expert systems in real-time applications reveals the necessity of reducing response times. Since monoprocessor optimizations of production systems have widely been explored, only multiple processor architectures appear to provide further performance gain. Efficient exploitation of the inherent parallelism of production systems, however, requires suitable algorithms for load balancing without simultaneously increasing organization or communication overhead. We present a novel parallel algorithm for PAMELA expert systems, based on dynamic distribution of data processing. The concept is supported by a transputer based architecture with an advanced interconnection structure.

Guaranteeing serializable results in synchronous parallel production systems

To speed up production systems, researchers have studied how to execute many rules simultaneously. Unfortunately, such systems can yield results that are impossible for a serial system to produce, leading to erroneous behaviors. We present a formal solution to the problem of guaranteeing serializable behavior in synchronous parallel production systems that execute many rules simultaneously. We also present a variety of algorithms that implement this solution. Our framework builds upon the work of Ishida and Stolfo [14] and improves upon their solution. The practical advantages of these strategies are demonstrated with measurements from an initial implementation.

Minimal state space search in parallel production systems

A methodology for transforming the original search space into an equivalent but minimal search space is proposed. First, the concept of dependences leads to a procedure for reduction of the search space. The search procedure using this method can produce a minimal and complete search space. It is shown that this method is applicable to parallel search as well. An added advantage of this method is that it does not exclude the use of heuristics. pi - lambda transformation is introduced to reduce the parallel search space.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Data-driven parallel production systems

Much effort has been expended on developing special architectures dedicated to the efficient execution of production systems. While data-flow principles of execution offer the promise of high programmability for numerical computations, it is shown that the data-driven principles can also be applied to symbolic computations. In particular, a mapping of the RETE match algorithm along the line of production systems is considered. Bottlenecks of the RETE match algorithm in a multiprocessor environment are identified and possible solutions are suggested. The modifications to the actor set as well as the program graph design are shown for execution on the tagged data-flow computer. The results of a deterministic simulation of this multiprocessor architecture demonstrate that artificial intelligence production systems can be efficiently mapped on data-driven architectures. >

The allocation problem in parallel production systems

Parallelism existing in a production system at different levels may be used to speed up the execution. In this paper, the notion of dependence between rules is introduced. It is shown that the analysis of dependences reveals the inherent parallelism and communication requirements among the rules. Techniques for automatic detection of parallelism and communication requirements are presented. A total parallel model which exploits parallelism in all the phases including matching and firing productions is assumed. The execution of a production of system on a multiprocessor involves the problem of partitioning the database and the rules among the processors. This is formulated as a 0–1 linear-quadratic programming problem and is shown to reduce to 0–1 linear programming with the penalty of increasing the size of the problem. Finally, a heuristic is given to solve the allocation problem using the A ∗ algorithm.

RUBIC: a multiprocessor for rule-based systems

It is shown how sequential production systems can be transformed into equivalent parallel forms by performing an analysis of rule interdependence. In the parallel production system model, rules fire simultaneously and the search space is reduced from the original form. A multiprocessor called RUBIC (rule-based inference computer) was designed to implement the parallel processing model. RUBIC has a message-passing architecture. The partitioning and mapping of production systems into the multiprocessor is achieved by optimizing a performance index such that inherent parallelism is maximized and interprocessor communication is minimized. >

Generative models in computer-assisted musical composition

A number of generative models have been used to facilitate the handling of large amounts of data in computer assisted musical composition. Various modelling techniques are identified and compared. These include fractal (self-similar) models used in both stochastic and non-stochastic contexts. A technique is explained which makes use of recursive parallel production systems similar to those which have been used to devise effective models of biological growth. By applying a cyclic sequence of generative rules to divide up the pitch-time-timbre space, complex non-stochastic musical structures may be formed which exhibit multi-level self-similar features. Minor modifications of an initial seeding cycle may produce a variety of different outcomes as in each successive generation the controlling cycle moves in and out of phase with the number of nodes available for further growth.

Skill Formation in Comparative Perspective: Malaysia and Japan

Abstract The concept of skill formation, which has emerged as critically important in post-Braverman labour process theory, is presented as a key concept in the developing debate on the specificity of Japanese management. As a test case for this, the paper presents empirical data on the comparison of skill levels and skill development in parallel production systems in Malaysia and Japan. The specific case study is of brewing, which is a traditional process industry.

A parallel rule firing fuzzy production system with resolution of memory conflicts by weak fuzzy monotonicity, applied to the classification of multiple objects characterized by multiple uncertain features

A fuzzy production system shell is described characterized by parallel rather than sequential rule firing. All fireable rules are fired in effect concurrently. Since there is no unfired-rule stack, no backtracking can take place, and no rule conflict algorithm is necessary; instead, a memory conflict algorithm is invoked when more than one rule modifies the same datum. Memory conflicts are resolved by weakly monotonie fuzzy logic; i.e. the value or truth value of an attribute may be replaced if the new truth value is equal to or greater than the old truth value. The system depends heavily on the use of fuzzy logic and on confidence levels, fuzzy numbers and fuzzy sets as explicit data types, and on the generation of rules from a data base of expert knowledge. Fuzzy sets are used to store contradictory and ambiguous information and results. If a problem is suitable for parallel processing, substantial reductions in system overhead are achieved, together with substantial economy in the number of rules which must be written; if a problem is not suitable for parallel processing, no economy is achieved. We suggest that problems which yield to deductive reasoning constitute a class which is suitable for sequential rule firing, and problems which yield to inductive reasoning constitute a class suitable for parallel processing. A successful application of the system to the unsupervised analysis of a time sequence of noisy echocardiogram images is described.

THE EMPLOYMENT EFFECTS OF 1980 PRICE AND WAGE POLICY IN THE ZIMBABWE MAIZE AND TOBACCO INDUSTRIES

ZIMBABWE, despite a fairly sophisticated modern sector, displays the features typical of developing countries with agricultural sectors characterized by parallel modern and traditional production systems and large surplus labour forces. Reynolds has observed that poor, labour-surplus countries are still living in a classical world and should perhaps follow the classical route towards fuller employment.' The reality, however, is that social and political factors tend to dominate classical economics particularly in democratic developing countries such as Zimbabwe, where wage-earners are highly influential. The government elected in Zimbabwe in 1980 has taken control of a nation with a strong commercial agricultural sector2 (predominantly white, owneroperator farmers) and a poorly developed peasant agricultural sector. The peasant sector has, for varying periods, suffered from a combination of antagonistic government policies, guerrilla warfare and rapid population growth. Zimbabwe is characterized by a high degree of food self-sufficiency, a wide gap between the upper and lower income-earners in the economy, and the need to maintain confidence among the many divergent groups which make up the nation. The improvement in income and living conditions for the poor of Zimbabwe is a priority task of the new government. At the same time, the need to sustain and develop existing commercial agriculture is recognized. A strong agricultural sector will have a major part to play in the development of both the country and the region in terms of food production and foreign exchange earnings. The development of the peasant sector will require massive inputs of resources in the immediate future while the commercial sector will be expected to maintain or increase production under less-than-ideal economic conditions. How to find the correct balance between improving the lot of both wage-earners and peasants while maintaining confidence in the commercial, industrial and agricultural sectors of the economy is a central dilemma currently facing Zimbabwe's new government.