Oussama Aoun

Hidden Markov Model Classifier for the Adaptive Particle Swarm Optimization

Particle swarm optimization (PSO) is a stochastic algorithm based population that integrates social interactions of animals in nature. Adaptive Particle swarm optimization (APSO) as an amelioration of the original one, improve the performance of global search and gives better efficiency. The APSO defines four evolutionary states: exploration, exploitation, convergence, and jumping out. According to the state, the inertia weight and acceleration coefficients are controlled. In this paper, we integrate Hidden Markov Model Particle swarm optimization (HMM) in APSO to have a stochastic state classification at each iteration. Furthermore, to tackle the problem of the dynamic environment during iterations, an additional online learning for HMM parameters is integrated into the algorithm using online Expectation-Maximization algorithm. We performed evaluations on ten benchmark functions to test the HMM integration inside APSO. Experimental results show that our proposed scheme outperforms other PSO variants in major cases regarding solution accuracy and specially convergence speed.

Application of multi-agent Markov decision processes to gate assignment problem

Gate assignment Problem (GAP) is an important subject of airport management to ensure smooth traffic operations. However, flights schedule may undergo some stochastic events such as delays that usually occur and have to be considered in the planning. Our approach considers the representation of gates as collaborative agents trying to complete a set of flights assignment tasks as given by a centralized controller. That will allow giving a new model for the GAP based on Multi Agent Markov Decision Processes (MMDP). The aim of this work is to give to controllers at the airport a robust priory solution instead of taking the risk of online schedule modifications to handle uncertainty. The solution of this problem will be a set of optimal decisions to be taken in every case of traffic disturbance.

Using Markov decision processes to solve stochastic gate assignment problem

Airport traffic often undergoes some random disruptions, which have to be considered to ensure operational flight-gate assignments. Inadequate assignment of gates may result in flight delays that happen in airport operations and must be taken into account. Here, we are using the original algorithm, based on Markov decision process (MDP) to solve the gate assignment problem (GAP) under uncertainty; we include stochastic parameters that depend on probabilities to express fluctuations in flight operations. The use of MDP for modeling will provide for airport controllers a robust solution for the GAP that takes in consideration possible flight delays. This paper gives the corresponding model, which includes strict constraints of the GAP and other soft constraints like choice preferences of gates. We give experimental results on a sample of real data to demonstrate the feasibility and efficiency of our approach.

A Probabilistic Finite State Machine Design of Particle Swarm Optimization

Nowadays, control is the main concern with emergent behaviours of multi-agent systems and state machine reasoning. This paper focuses on the restriction of this general issue to swarm intelligence approaches designed for solving complex optimization problems. Indeed, we propose a probabilistic finite state machine for controlling particles behaviour of the particle swarm optimization algorithm. That is, our multi-agent approach consists of assigning different roles to each particle based on its probabilistic finite state machine control which is used to address this issue. We performed evaluations on ten benchmark functions to test our control scheme for particles. Experimental results show that our proposed scheme gives a distinguishable out-performance on a number of state of the art of PSO variants.

Self Inertia Weight Adaptation for the Particle Swarm Optimization

Particle swarm optimization is a stochastic population-based metaheuristic algorithm, it been successful in solving a height range of real-world problems. The primary challenge present in PSO is to balance between global and local search during the optimization process. The classical PSO algorithm applies the same search scheme for all particles, which implies that all the swarm uses the same strategy. This monotonic learning design produces the be short of learning ability for a single particle that makes the classical PSO unable to deal with diverse and complex optimization problems. In this paper, we present a new learning design for particles in the PSO algorithm. In the proposed approach, each particle balances between four learning strategies to influence and adapt its inertia weight. The self-adaptation is controlled by a Hidden Markov Chain of each particles states, which allows incorporating both the particle history and the current situation of the search process in the individual learning level of the swarm. The experimental study on a set of several benchmark functions shows remarkable performances in comparison with other PSO variances from literature.

Time-Dependence in Multi-Agent MDP Applied to Gate Assignment Problem

Many disturbances can impact gate assignments in daily operations of an airport. Gate Assignment Problem (GAP) is the main task of an airport to ensure smooth flight-to-Gate assignment managing all disturbances. Or, flights schedule often undergoes some unplanned disruptions, such as weather conditions, gate availability or simply a delay that usually arises. A good plan to GAP should manage as possible stochastic events and include all in the planning of assignment. To build a robust model taking in account eventual planning disorder, a dynamic stochastic vision based on Markov Decision Process theory is designed. In this approach, gates are perceived as collaborative agents seeking to accomplish a specific set of flights assignment tasks as provided by a centralized controller. Multi-agent reasoning is then coupled with time dependence aptitude with both time-dependent action durations and stochastic state transitions. This reflection will enable setting up a new model for the GAP powered by a Time-dependent Multi-Agent Markov Decision Processes (TMMDP). The use of this model can provide to controllers at the airport a robust prior solution in every time sequence rather than bringing a risk of online schedule adjustments to handle uncertainty. The solution of this model is a set of optimal decisions time valuated to be made in each case of traffic disruption and at every moment.

Data-driven based aircraft maintenance routing by markov decision process model

Aircraft maintenance routing is of basic significance to the safe and efficient operations of an airline. However, the timely efficiency of the airline flight schedule is susceptible to various factors during the daily operations. Air traffic often undergoes some random disruptions that expose maintenance routing to random flight delays, which have to be considered to ensure safe and operational flight schedule. The idea of data-driven methods was the focal point of much studies during a previous couple of years. Constrained Markov Decision process model was selected in this paper to remedy this problem and design the maintenance needs of an aircraft taking past data information into account. Maintenance actions are so modeled with stochastic state transitions. This can offer the opportunity to solve the maintenance routing problem deliberating and handling flight disturbances. Through computational tests on real data of a Moroccan airline company, we investigate the efficiency of this solution approach on history data sets.