Bernard Archimède

Towards a distributed multi-agent framework for shared resources scheduling

Nowadays, manufacturers have to share some of their resources with partners due to the competitive economic environment. The management of the availability periods of shared resources causes a problem because it is achieved by the scheduling systems, which assume a local environment where all resources are on the same site. Therefore, distributed scheduling with shared resources is an important research topic. In this communication, we introduce the architecture and behaviour of DSCEP framework (Distributed, Supervisor, Customer, Environment, and Producer) under shared resources situation with disturbances. We are using a simple example of manufacturing system to illustrate the ability of DSCEP framework to solve the shared resources scheduling problem in complex systems.

A flood lamination strategy based on transportation network with time delay

Over the last few years, the frequency and intensity of floods has become more marked due to the influence of climate change. The engendered problems are related to the safety of goods and persons. These considerations require predictive management that will limit water height downstream. In the literature, numerous works have described flow modeling and management. The work presented in this paper is interested in quantitative management by means of flood expansion areas placed along the river and for which we have size and location. The performance of the management system depends on the time and height of gate opening, which will influence wave mitigation. The proposed management method is based on use of a transportation network with time delay from which the volume of water to be stored is calculated.

Towards a distributed framework for transportation planning: A food supply chain case study

Distribution and remoteness of production sites of enterprise networks, remoteness and multiplicity of distribution centers, the explosion of e-commerce has led to an increase in the number of requests for transportation around the world. This increase in the volume of transport of goods and cargo added to the growing number of passenger trips has led to an increase in transportation means(cars, planes, boats, etc.), with a consequent increase in the capacity of communication channels, which reached saturation (highways, airlines, shipping lanes), an expansion of storage areas (ports, airports, warehouses, etc.), and an increase in pollution impacting the environment sustainability. In this context, organization, management and transportation planning, become crucial, favored the emergence of many specialized companies (3PL) offering a pooling of transport and centralized management. The objective of this paper is to present a distributed architecture planning of transportation activities aimed at better utilize transport resources by grouping several orders of transport for each effective displacement.

A multi-agent distributed framework for collaborative transportation planning

Abstract Today enterprises are organized in networks of several distributed production sites in a geographical area more and more extended, where sites use the intermediary products of other sites to manufacture the final products. Therefore they transport raw materials and intermediate products between plants for manufacturing. Additionally these enterprises have to distribute their final products to far away consumers. In this context, it is difficult and not cost effective for these enterprises to manage their own transport vehicles. Thus they outsource their transportation tasks to third party logistics enterprises. These enterprises group transport orders of multiple enterprises together in order to reduce cost and travels number. The objective of this paper is to present a distributed architecture planning of transportation activities aimed at better utilize transport resources by grouping several orders of transport for each effective displacement.

A distributed multi-agent framework for shared resources scheduling

Nowadays, manufacturers have to share some of their resources with partners due to the competitive economic environment. The management of the availability periods of shared resources causes a problem because it is achieved by the scheduling systems which assume a local environment where all resources are on the same site. Therefore, distributed scheduling with shared resources is an important research topic in recent years. In this communication, we introduce the architecture and behavior of DSCEP framework (distributed, supervisor, customer, environment, and producer) under shared resources situation with disturbances. We are using a simple example of manufacturing system to illustrate the ability of DSCEP framework to solve the shared resources scheduling problem in complex systems.

Application of using SCEP model for distributed scheduling with shared resources in hospital system

The increasing intense of the economic constraints forces the manufacturers to share their resources with other organizations [1]. But the management of the available periods of shared resources causes a problem because it is mostly realized in a centralized way [2]. Therefore, distributed scheduling with shared resources is nowadays an important research topic. In this paper, we present a multi-agent model named SCEP (Supervisor, Customer, Environment and Producer). We also introduce the detailed architecture and behavior of the SCEP model in the context of multi-site and shared resources situation. The SCEP application proposed here, aims at improving the scheduling of shared resources in a hospital framework.

A water storage and release strategy for flood management based on transportation network with time delay

Today, with frequency and severity of extreme weather, flood management became in the world a major concern for governments. The engendered problems are relative to goods and persons safety. Their considerations require a predictive management involving storage and draw-off strategy with several flood areas in order to limit water heights to the downstream of streams. A good flood management requires a control not only of storage to avoid downstream problems but also of water release to protect the fertility of flood expansion areas which are most often based on growing areas. In the literature, numerous described works were done on flows modeling and management. The work presented in this paper, is interested in the quantitative management by means of floods expansion areas placed along the river and for which we have their location as well as their sizing. The performances of management depend on moment and on height of gates opening which influence on wave mitigation. The proposed management method is based on a dynamic mechanism from which the water volumes to be stored or draw-off in time are calculated.

Interoperability Service Utility Model and its Simulation for Improving the Business Process Collaboration

In this paper an interoperability service utility (ISU) model is defined and formalized for improving the collaboration between partners. The ISU model makes it possible to limit the associated risk during the exchange of herogeneous information between enterprises. Another advantage of the approach consists in the simulation-based evaluation method of the impact of interoperability parameters on the performances of the business process. The integration of the ISU model was achieved at the business activity level and a method for performance evaluation is presented. The effectiveness of the ISU model was studied in a national industrial project and is shown herein on a simple example of P2P collaboration.

Combining multi-agent model, SOA and ontologies in a distributed and interoperable architecture to manage multi-site production projects

Today, production projects are often achieved by multi-site production and the associated markets become more and more dynamic and service oriented. In this context, the multi-site project management is a difficult task, which has to take into account not only the decision distribution and the information confidentiality, but also the heterogeneity existing between different production management systems used by partners. The generic planning model SCEP (supervisor, customer, environment, producer) based on multi-agent technology offers a well-adapted solution to the decision distribution and information confidentiality in project management. SCEP was integrated with ontological mechanisms and SOA (service-oriented architecture) components in a new architecture Semantic-SCEPSOA, which makes it possible to achieve a distributed and interoperable management of multi-site production projects. In the case of two sites managed by two different planning software applications, the semantic interoperability effectiveness is shown within the S-SCEPSOA architecture.

Semantic Interoperability in Astrophysics for Workflows Extraction from Heterogeneous Services

Modern instruments in astrophysics lead to a growing amount of data and more and more specific observations, among which scientists must be able to identify and retrieve useful information for their own specific research. The Virtual Observatory (http://www.ivoa.net/deployers/intro_to_vo_concepts.html) architecture has been designed to achieve this goal. It allows the joint use of data taken from different instruments. Retrieving and cross-matching those data is in progress, but it’s impossible today to find a sequence resolving a given science case needing a combination of existing services of whom the user doesn’t knows the specifications. The goal of this work is to propose the basis of an architecture leading to automatic composition of workflows that implement scientific use cases.