Alexandre Sava

Simulation-Based Discrete Optimization of Stochastic Discrete Event Systems Subject to Non Closed-Form Constraints

This technical note addresses the discrete optimization of stochastic discrete event systems for which both the performance function and the constraint function are not known but can be evaluated by simulation and the solution space is either finite or unbounded. Our method is based on random search in a neighborhood structure called the most promising area proposed in and a moving observation area. The simulation budget is allocated dynamically to promising solutions. Simulation-based constraints are taken into account in an augmented performance function via an increasing penalty factor. We prove that under some assumptions, the algorithm converges with probability 1 to a set of true local optimal solutions. These assumptions are restrictive and difficult to verify but we hope that the encouraging numerical results would motivate future research exploiting ideas of this technical note.

An analytical approach for performance evaluation and optimization of a two-stage production-distribution system

This paper addresses the performance evaluation and optimization of a production-distribution system composed of a warehouse supplied by a manufacturing plant with finite capacity. Customer orders arrive randomly with random order size. Transportation time from plant to warehouse is fixed. We propose an analytical approach that uses only the first two moments of random variables to evaluate the inventory cost and demand fill rate. A gradient-based method is then proposed to minimize the average total inventory cost subject to fill rate requirement. Numerical comparisons with simulation show that the analytical approach is very efficient.

Assembly Line Balancing Problems Solved by Estimation of Distribution

In this paper, we propose an new algorithm based on an estimation of distribution (ED) to solve the type II assembly line balancing problem (ALBP-II). This problem aims to assign a set of assembly operations subject to precedence constraints to a given number of workstations of an assembly line in order to minimize the cycle time. We prove that the optimal solution for determinist ALBP-II problem can maximize the reliability of the least reliable station in the line if the operation times are random and normally distributed with constant variance-to-mean ratio. The ED algorithm is a population-based algorithm. The simulation results show that ED algorithm outperforms the simulated annealing algorithm especially for large-scaled problems.

Hybrid Petri nets for modeling and control of multi-source energy conversion systems

This paper proposes an approach of hybrid Petri nets modeling for hybrid renewable energy production systems structured in micro-grids. This approach is original in the sense that it enables to study the behavior of such systems under various reconfiguration constraints in order to fulfill energy demands for customers. The proposed Petri net model allows to considering the discreet events related to the availability of energy resources or to components failures in the system as well in the discreet reconfiguration (structural or functional), for control strategies specifications. In addition, an operational research model for cost minimization and availability maximization through reconfiguration is also proposed. The joint implementation of both models should lead to a new approach of design of hybrid renewable energy systems and of their effective cost optimization.

Supervisory Control of partially observable Marked Graph based on Marking Exclusion Constraint

This paper presents a control synthesis approach for discrete event systems modelled by partially observable marked graphs. It solves the forbidden state problem for discrete event systems modelled by partially observable event graphs subject to Marking Exclusion Constraints (MEC). Two kind of MEC specifications are considered: MEC-AND or MEC-OR. The main advantage of MEC specifications is an increased modelling power regarding to General Mutual Exclusion Constraints GMEC. We provide a simple optimal controller to enforce MEC specifications with independent critical places. We also propose a suboptimal controller for MEC-OR with dependent critical places belonging to the same direct path.

Control Synthesis for Partially Controlled Marked Graph Subject to Marking Exclusion Constraints

The goal of control synthesis is to build a control law which forces the system to respect given specifications. In this paper we propose a new type of specification called Marking Exclusion Constraint (MEC). The main advantage of MEC specification is an increased modeling power regarding General Mutual Exclusion Constraints (GMEC). We define two types of MEC: MEC-OR and MEC-AND and we propose a technique to build the controller which enforces MEC specifications for discrete events systems modeled by marking graphs.

Reconfigurable manufacturing system: Overview and proposition of new approach

To confront the rapid change in manufacturing environments, companies need a Reconfigurable Manufacturing System (RMS) which combines the high production volume of Dedicated Manufacturing Lines (DML) with the large variety of products of Flexible Manufacturing System (FMS). In this paper, we propose a state of art of recent research work on RMS. After highlighting the advantages of RMS compared to existing manufacturing systems, we outline different issues discussed in the context of RMSs and what it brings as solutions. Finally we present the main idea of a new approach for reconfiguration process following a predictive monitoring.

Monte Carlo based Petri net simulation for maintenance strategies assessment in series-parallel-series multi-physic systems

The authors propose a methodology to assess the effectiveness of a maintenance strategy on the availability of a serial-parallel multi-physic system, using Monte Carlo simulation embedded in a Petri net model. The systems are composed of heterogenous components that are characterized by specific degradations and failure mechanisms. Building an effective maintenance strategy to improve the availability of such a system requires to monitoring the degradation of each component. We assume that each component is subject to stochastic degradations. Also, we consider that each component might have three health status, according to degradation thresholds, function of the component reliability: “healthy”, “degraded” and “failed”. The health condition of the overall system relies on the health status of each component. A model for tracking the status of each component has been worked out using a colored stochastic Petri net (CSPN). Each health status is modeled by a place within the CSPN model, where each component is modeled by a colored token. The degradation of each component of the system is evaluated based on the Monte Carlo simulation technique. Transition firing regarding a given color model the evolution of the associated component from a health condition to another due to the degradation mechanism or to a maintenance action aimed to restore partially or totally its performance. However, the degradation of each component does not have the same influence on the performance of the overall system. Operational performance indicators are introduced to quantify the influence of each component on the performance of the entire system. Furthermore, maintenance actions are defined taking into account the degradation level of each component, the influence that each component has on the performance of the system and the available repairman. The effectiveness of the maintenance strategy on the system availability is evaluated through simulation.

ISO 9001 International Standard, a Tool to Enhance Data Quality in Durable Socio-Technical Systems

The quality of data is recognized to be a key issue for the assets management in enterprises as data is the foundation of any decision making process. Recent research work has established that the quality of data is highly dependent on the knowledge one has on the socio-technical system being considered. Three modes of knowledge have been identified: knowing what, knowing how and knowing why. In this paper we focus on how to manage these modes of knowledge in durable socio-technical systems to enhance the data quality face to technological progress and employees turnover. We believe that an organization based on ISO 9001 international standard can provide a valuable framework to provide the data quality needed to an efficient decision making process. This framework has been applied to design the data quality management system within a high education socio-technical system. The most important benefits that have been noticed are: 1) a shared vision on the external clients of the system with a positive impact on the definition of the strategy and the objectives of the system and 2) a deep understanding of the data client-supplier relationship inside the socio-technical system. A direct consequence of these achievements was the increasing knowledge on “know-what” data to collect, “know-why” to collect that data and “know-how” to collect it.

Internal Energy Use in Calculating the Lifetime of 2P Armor Steel

Armored steel sheets have various uses in civil and military industry. For the last one, especially in aeronautics, a better determination of the lifetime for the fatigue and shock loaded parts, is a major challenge. Several methods for fatigue calculus are known: safe-life, fail-safe, crack propagation method. All of this methods are not considering in any way the shocks that can accidentally occur, so in the calculation of lifetime, the role of impact multiplier is null. The authors propose a corrected formula for the calculus of the lifetime for 2P armor steel, based on the internal energy developed into the test specimens, through the impact multiplier.