Saoussen Krichen

A multi-objective optimization approach for resource assignment and task scheduling problem: Application to maritime domain awareness

Large volume surveillance missions are characterized by the employment of mobile and fixed surveillance assets to a large geographic operation area in order to perform surveillance activities. Finding efficient management solutions should be investigated to optimize assets allocation and tasks achievement. In this paper, we propose to model this optimization problem as a multi-objective, multi-mode assignment and scheduling problem. Resources are to be assigned to accomplish the tasks. Then, surveillance tasks should be scheduled onto successive periods. The problem is designed to consider two conflicting objective functions: minimizing the makespan and minimizing the total cost. As the problem is NP-Hard, a bi-colony ant based approach is proposed. The empirical validation is done using a simulation environment Inform Lab. The experimental results show that the computational time remains polynomial with respect to the problems size.

A genetic algorithm for a multi-objective nodes placement problem in heterogeneous network infrastructure for surveillance applications

In this paper, we adress a Multi-objective communication nodes (e.g., antennas, relays…) placement problem for heterogeneous network infrastructure. The proposed model considers three conflicting objective functions: maximizing the communication coverage, minimizing the cost of nodes placement and communication devices and the maximizing of the total capacity bandwidth in the network. The empirical validation of the model is done in a simulation environment called “Inform Lab”. We consider a large volume of surveillance missions. To solve such an NP-Hard problem, we propose a Multi-objective Genetic Algorithm (MOGA). The empirical results show that the proposed algorithm has good performance with good qualitys result in a practicable CPU time.

Geographical information systems and spatial optimization

Introduction Geographical Information Systems: Basic Concepts Introduction Geographical databases Geographical information systems Research areas Conclusion Optimization: Basic Concepts Introduction Design of an optimization problem Features of an optimization problem Potential problems in optimization Solution approaches Conclusion Integration Strategies of GIS and Optimization Systems Introduction The importance of GIS-O integration strategies The full GIS-O integration strategy The loose GIS-O integration strategy The tight GIS-O integration strategy Comparison between integration strategies Potential applications of GIS-based optimization tools Conclusion A GIS-O Framework for the Vector Loading Distance Capacitated Vehicle Routing Problem Introduction General context VRP variants The VL-DCVRP A loose GIS-O integration for the VL-DCVRP Conclusion References

An ant colony optimization metaheuristic for solving bi-objective multi-sources multicommodity communication flow problem

This paper studies the single path multi-sources multicommodity communication flow problem (MMCF). A predefined number of messages are to be routed in a capacitated network including a set of nodes that can be producers (sources) and/or consumers (destinations, decision makers) of information. A node might also be a simple relay. We assume that the same information might be provided by different sources. Each edge in the network is characterized by a capacity, a transmission delay and a cost. We propose a mathematical formulation of the MMCF as a biobjective optimization problem that minimizes the overall cost and delay. Network structural constraints are to be respected such as the capacity of the edges and the single path. We assume the non preemptiveness of the transmission. A solution of the proposed model provides for each request the assigned source node, the transmission path as well as the bandwidth allocated along the path. Multicast trees might be generated if the same source is assigned to send the same message to different destinations. An ant colony metaheuristic is proposed to solve the problem. This paper presents an empirical validation of the proposed approach.

An Adaptive Tabu Search Algorithm for the Multi-Objective Node Placement Problem In Heterogeneous Networks

The Multiâx80x93objective Node Placement (MONP) problem focuses on extending an existing communication in- n nfrastructure with new wireless heterogeneous network components while achieving cost effectiveness and ease n nof management. This extention aims to broaden the coverage and handle demand fluctuations. In this paper, n nthe MONP problem is modelled as a multiâx80x93objective optimization problem with three objectives: maximiz- n ning the communication coverage, minimizing active nodes and communication devices costs, and maximizing n nof the total capacity bandwidth in the network. As the MONP problem is N Pâx80x93Hard, we present a metaâx80x93 n nheuristic based on the Tabu Search approach specifically designed for multiâx80x93objective problems in wireless n nnetworks. We present an empirical validation of the model and the algorithm based on a selection of a real n nand large set of instances. We present also a performance comparison between the suggested algorithm and a n nmultiâx80x93objective genetic algorithm (MOGA). All tests are performed on a real simulation environment for the n nmaritime surveillance application.

A multiobjective hybrid ant colony optimization approach applied to the assignment and scheduling problem

The assignment and scheduling problem is inherently multiobjective. It generally involves multiple conflicting objectives and large and highly complex search spaces. The problem allows the determination of an efficient allocation of a set of limited and shared resources to perform tasks, and an efficient arrangement scheme of a set of tasks over time, while fulfilling spatiotemporal constraints. The main objective is to minimize the project makespan as well as the total cost. Finding a good approximation set is the result of trade-offs between diversity of solutions and convergence toward the Pareto-optimal front. It is difficult to achieve such a balance with NP-hard problems. In this respect, and in order to efficiently explore the search space, a hybrid bidirectional ant-based approach is proposed in this paper, which is an improvement of a bi-colony ant-based approach. Its main characteristic is that it combines a solution construction developed for a more complicated problem with a Pareto-guided local search engine.

Particle swarm optimization approach for resolving the cutting stock problem

We propose in this paper a one dimensional cutting stock problem encountered in a large manufacturer of multi-usable cables. The problem consists in the post-production process where a variety of orders of multi-sized sets of cables should be satisfied. We developed a solution method based on a particle swarm optimization approach that takes into account the characteristics of the specific problem. We assume that the manufacture produce cables sets having the same length. The reduction of the wastage is tackled which is successfully tested in this paper. A mathematical model of the problem is developed, and results on a wide variety of instances and comparisons with other works found in the literature are presented to illustrate the effectiveness of our algorithm in solving the cutting stock problem.

A particle swarm optimization for solving the one dimensional container loading problem

We address in this paper the one dimensional container loading problem (CLP), a NP-hard optimization problem of extreme economic relevance in industrial areas. The problem consists in loading items into containers, then stowing the most profitable containers in a set of compartments. The main objective is to minimize the number of used containers. We state a mathematical model as well as a modified metaheuristic namely the particle swarm optimization approach (PSO) with FFD initialization. Computational results carried out on a large test bed show the effectiveness of the denoted approach depending on the problem settings.

A hybrid genetic approach for multi-objective and multi-platform large volume surveillance problem

Efficient management of surveillance assets and successful scheduling of surveillance tasks are complex decision-making problems for the execution of large volume surveillance missions in order to improve security and safety. A mission can be seen as a defined set of logical ordered tasks with time and space constraints. The resources to task assignment rules require that available assets should be allocated to each task. A combination of assets might be required to execute a given task. Finding efficient management solutions should be investigated to optimise assets-resources allocation and tasks scheduling. In this paper, we propose to model this optimisation problem as a multi-objective, multi-platform assignment and scheduling problem. Resources are to be assigned to accomplish different tasks. Surveillance tasks should be scheduled into successive periods. The problem is designed to consider two conflicting objective functions: minimising the makespan and minimising the total cost. As the problem is NP-hard, a hybrid genetic algorithm HGA is proposed. The empirical validation is performed using a simulation environment called Inform Lab, and a comparison to two state-of-the-art multi-objective approaches based on selected performance metrics. The experimental results show that HGA performs consistently well for high dimensional problems.

A Multi-start Tabu Search Approach for Solving the Information Routing Problem

In this paper, we propose a new global routing algorithm supporting advance reservation. A set of flows are shared across a communication network. Each flow has a source node, a destination node and a predetermined traffic demand. The design goal of the routing is to minimize the overall network congestion under the constraint that each flow should be sent along a single path without being bifurcated. We model this optimization problem as a single path multicommodity flow problem (SPMFP). As the complexity of the SPMFP is NP-Hard, a Multi-start Tabu Search (MTS) is proposed as a solution approach. The empirical validation is done using a simulation environment called Inform Lab. A comparison to a state-of-the-art ant colony system (ACS) approach is performed based on a real case of maritime surveillance application. The same instances are optimally solved using CPLEX. The experimental results show that the MTS produces considerably better results than the ACS to the detriment of the CPU time.