Abdelhakim Artiba

A hybrid three-stage flowshop problem: Efficient heuristics to minimize makespan

We treat a problem of scheduling n jobs on a three stages hybrid flowshop of particular structure (one machine in the first and third stages and two dedicated machines in stage two). The objective is to minimize the makespan. This problem is NP-complete. We propose two heuristic procedures to cope with realistic problems. Extensive experimentation with various problem sizes are conducted and the computational results show excellent performance of the proposed heuristics.

Manufacturing execution system – a literature review

Since their beginning in the mid 1990s, Manufacturing Execution Systems (MES) have significantly evolved into more powerful and more integrated software applications as computing technologies have advanced. Their functionality coverage has changed significantly and can now provide a common and single system to support most of the manufacturing execution processes from the production order release to the delivery of finished goods. However, MES applications still fall short of adding capabilities for decision-making tools in very dynamic organisations where adaptive execution strategies are required to replenish the supply chain while dynamically responding to unpredicted change. This article intends to describe what MES have become, present their relationships with other enterprise information systems, and to identify major issues related to MES. We also discuss research areas that need to be explored in order to resolve the increased complexity of execution systems and to fulfil the continuing customer needs for faster real-time response and expanded functionality coverage.

Sequencing a hybrid two-stage flowshop with dedicated machines

We treat the n -job, two-stage hybrid flowshop problem with one machine in the first stage and two different machines in parallel in the second stage. The objective is to minimize the makespan. We demonstrate that the problem is NP-complete. We formulate a dynamic program, which is beyond our grasp for problems of more than 15 jobs. Our search for heuristic approaches led to the adoption of the Johnson sequence, which motivated two of the three approaches: dynamic programming and sequence-and-merge. The third approach, the greedy heuristic, was included as example of an elementary heuristic.

An application of a planning and scheduling multi-model approach in the chemical industry

In this paper, we are concerned with the production planning and scheduling of process industries. A real life application stems from the chemical industry where the production environment is subject to uncertainty, delays and interferences. We propose a multi-model based system for planning and scheduling in such a process industry. The multi-model system is defined as an architecture integrating expert system techniques, discrete event simulation, optimization algorithms and heuristics to support decision-making for complex production planning and scheduling problems. The system has been successfully tested and used to help in production planning and scheduling.

Minimizing the weighted sum of maximum earliness and maximum tardiness costs on a single machine with periodic preventive maintenance

We consider the problem of scheduling a set of jobs on a single machine against a common and restrictive due date. In particular, we are interested in the problem of minimizing the weighted sum of maximum earliness and maximum tardiness costs. This kind of objective function is related to the just-in-time environment where penalties, such as storage cost and additional charges for late delivery, should be avoided. First we present a mixed integer linear model for the problem without availability constraints and we prove that this model can be reduced to a polynomial-time model. Secondly, we suppose that the machine undergoes a periodic preventive maintenance. We present then a second mixed integer linear model to solve the problem to optimality. Although the latter problem can be solved to optimality for small instances, we show that the problem reduces to the one-dimensional bin packing problem. Computational results show that the proposed algorithm best fit decreasing performs well.

Stochastic single machine scheduling with random common due date

This paper studies the single machine scheduling problem for minimising the expected total weighted deviations of completion times from random common due date. Jobs have exponentially distributed processing times and the common due date is a generalised Erlang distribution. The optimal schedules are shown to be ∧-shaped. Moreover, we give the optimal schedules when the machine is subject to stochastic breakdowns with independent and exponentially distributed uptimes and downtimes.

Simulation‐based approach to joint production and preventive maintenance scheduling on a failure‐prone machine

Purpose – The purpose of this study is to propose and model an integrated production‐maintenance strategy for a failure‐prone machine in a just‐in‐time context.Design/methodology/approach – The proposed integrated policy is defined and a simulation model is developed to investigate it.Findings – The paper focuses on finding simultaneously two decision variables: the period (T) at which preventive maintenance actions have to be performed; and the sequence of jobs (S). These values minimize the maintenance costs (MC) and the expected total earliness and tardiness costs (ETC) away from a common due‐date D.Practical implications – The paper attempts to integrate in a single model the two main aspects of any manufacturing and production systems: production and maintenance. It focuses on a stochastic scheduling problem in which n immediately available jobs are to be scheduled jointly with the preventive maintenance. The effect of the period (T) and the sequence of job (S) on the expected total cost are shown th...

Nested general variable neighborhood search for the periodic maintenance problem

In this paper we study the periodic maintenance problem: given a set of m machines and a horizon of T periods, find indefinitely repeating itself maintenance schedule such that at most one machine can be serviced at each period. In addition, all the machines must be serviced at least once for any cycle. In each period the machine i generates a servicing cost bi or an operating cost which depends on the last period in which i was serviced. The operating cost of each machine i in a period equals ai times the number of periods since the last servicing of that machine. The main objective is to find a cyclic maintenance schedule of a periodicity T that minimizes total cost. To solve this problem we propose a new Mixed Integer programming formulation and a new heuristic method based on general Variable neighborhood search called Nested general variable neighborhood search. The performance of this heuristic is shown through an extensive experimentation on a diverse set of problem instances.

Hybrid Flow Shop Scheduling with Availability Constraints

The success of a product depends on the costs incurred through its entire processing. Indeed, an efficient schedule can significantly reduce the total costs. Most of the literature on scheduling assumes that machines are always available. However, due to maintenance activities machines cannot operate continuously without some unavailability periods. This chapter deals with scheduling a hybrid flow shop with availability constraints to minimize makespan. We investigate exact methods to solve two special cases of this problem to optimality. We formulate a dynamic programming to solve two-machine flow shop and a branch and bound algorithm to solve the two-stage hybrid flow shop.

On the single-processor scheduling problem with time restrictions

This paper presents a time-indexed mixed integer programming formulation for the single-processor scheduling problem with time restrictions that has been formulated at first by Braun et al. in [1]. The problem is as follows. A set of n independent jobs are simultaneously available for processing at the beginning of the planning horizon, and their processing times are fixed and known in advance. The jobs have to be processed non-preemptively on a single processor that can handle only one job at a time. Furthermore, during any time period of length α > 0 the number of jobs being executed is less than or equal to a given integer value B ≥ 2. The objective is to minimize the completion time of the last job in the optimal sequence (i.e. the makespan). To our knowledge, this is the first time that a mathematical model is given to solve the single-processor scheduling problem with time restrictions exactly. The performance of the model is tested by running it on randomly generated instances. The computational analysis shows that the proposed model, without any valid cuts, performs considerably well for small instances and a relatively large value of the integer B.