Mark P. Kleeman

Solving the aircraft engine maintenance scheduling problem using a multi-objective evolutionary algorithm

Scheduling problems are a very common research topic. This is because, for efficiency reasons, our world relies heavily on schedules and deadlines. Aircraft engine maintenance is no exception. The United States Air Force has many planes that it must keep up and running. But with the downsizing that has occurred in recent years, the number of planes that are operational has become more critical. This means that every effort needs to be made to ensure that not only are the engines repaired in an efficient manner, but that their components scheduled maintenance cycles are in sync so that the engine has fewer trips to the logistics maintenance center.

Solving the multi-objective quadratic assignment problem using a fast messy genetic algorithm

The multiobjective quadratic assignment problem is an NP-complete problem with a multitude of real-world applications. The specific application addressed is the minimization of communication flows in a heterogenous mix of unmanned aerial vehicles. Developed is a multiobjective approach to solving the general mQAP for this UAV application. The combinatoric nature of this problem calls for a stochastic search algorithm; moreover, the multiobjective fast messy genetic algorithm (MOMGA-II) [Jesse Zydallis (2003)] is used for experimentation. Results indicate that much of the Pareto optimal points are found.

Scheduling of Flow-Shop, Job-Shop, and Combined Scheduling Problems using MOEAs with Fixed and Variable Length Chromosomes

This chapter introduces the novel multi-component schedul- ing problem, which is a combination of the generic flow-shop and job-shop (or open-shop) problems. This chapter first presents an overview of five common scheduling models and examples of how they are solved. A de- scription of som eo f themost common chromosome representations and genetic operators is also presented. The chapter also discusses som eo f the real-world problem st hat can bemodelled using the proposed multi- component scheduling model. In particular, the multi-component engine maintenance scheduling problem is presented and solved using a multi- objective evolutionary algorithm (MOEA) called GENMOP. A variable length chromosome is used by the MOEA in order to address problem specific and generic characteristics. The experimental results compare favorably to baseline values, indicating that GENMOP can effectively solve multi-component scheduling problems. Overall, this chapter intro- duces a new category of scheduling problem st hat is quite co mmon in real world problems and presents an example of the problem .B y intro- ducing this new category, which can have peculiarities that differ from other scheduling categories, researchers can build upon work done by others in this field.

Solving Multicommodity Capacitated Network Design Problems Using Multiobjective Evolutionary Algorithms

Evolutionary algorithms have been applied to a variety of network flow problems with acceptable results. In this research, a multiobjective evolutionary algorithm (MOEA) is used to solve a variation of the multicommodity capacitated network design problem (MCNDP). This variation represents a hybrid communication network as found in network centric models with multiple objectives including costs, delays, robustness, vulnerability, and reliability. Nodes in such centric systems can have multiple and varying link capacities, rates and information (commodity) quantities to be delivered and received. Each commodity can have an independent prioritized bandwidth requirement as well. Insight to the MCNDP problem domain and Pareto structure is developed. The nondominated sorting genetic algorithm (NSGA-II) is modified and extended to solve such a MCNDP. Since the MCNDP is highly constrained, a novel initialization procedure and mutation method are also integrated into this MOEA. Empirical results and analysis indicate that effective solutions are generated very efficiently

Solving Multicommodity Capacitated Network Design Problems using a Multiobjective Evolutionary Algorithm

Evolutionary algorithms can be applied to a variety of constrained network communication problems with centric type models. This paper shows that with real-world complex network communication problems of this type, sophisticated statistical search is required. This situation occurs due to the fact that these optimization problems are at least NP-complete. In order to appreciate the formal modeling of realistic communication networks, historical network design problems are presented and evolved into more complex real-world models with associated deterministic and stochastic solution approaches discussed. This discussion leads into the design of an innovative multiobjective evolutionary algorithm (MOEA) to solve a very complex network design problem variation called the multicommodity capacitated network design problem (MCNDP). This variation represents a hybrid real-world communication architecture as reflected in real-world network centric models with directional communications, multiple objectives including costs, delays, robustness, vulnerability, and operating reliability within network constraints. Nodes in such centric systems can have multiple and varying link capacities as well as rates and information (commodity) quantities to be sent and received. Each commodity can also have independent prioritized bandwidth and spectrum requirements. The nondominated sorting genetic algorithm (NSGA-II) is selected as the MOEA framework which is modified and parallelized to solve the generic MCNDP. Since the MCNDP is highly constrained but with an enormous number of possible network communication topologies, a novel initialization procedure and mutation method are integrated resulting in reduced search space. Empirical results and analysis indicate that effective topological Pareto solutions are generated for use in highly constrained, communication-based network design.

Multi-objective fast messy genetic algorithm solving deception problems

Deception problems are among the hardest problems to solve using ordinary genetic algorithms. Recent studies show that Bayesian optimization can help in solving these problems. This work compares the results acquired from the multiobjective fast messy genetic algorithm (MOMGA-II), multiobjective Bayesian optimization algorithm (mBOA), and the nondominated sorting genetic algorithm-II (NSGA-II) when applied to three different deception problems. The three deceptive problems studies are: interleaved minimal deceptive problem, interleaved 5-bit trap function, and the interleaved 6-bit bipolar function. The unmodified MOMGA-II, by design, explicitly learns building block linkages which is required if an algorithm is to solve these hard deception problems. Preliminary results using the MOMGA-II are favorable.

A multi-tiered memetic multiobjective evolutionary algorithm for the design of quantum cascade lasers

Recent advances in quantum cascade lasers (QCLs) have enabled their use as (tunable) emission sources for chemical and biological spectroscopy, as well as allowed their demonstration in applications in medical diagnostics and potential homeland security systems. Finding the optimal design solution can be challenging, especially for lasers that operate in the terahertz region. The production process is prohibitive, so an optimization algorithm is needed to find high quality QCL designs. Past research attempts using multiobjective evolutionary algorithms (MOEAs) have found good solutions, but lacked a local search element that could enable them to find more effective solutions. This research looks at two memetic MOEAs that use a neighborhood search. Our baseline memetic MOEA used a simple neighborhood search, which is similar to other MOEA neighborhood searches found in the literature. Alternatively, our innovative multi-tiered memetic MOEA uses problem domain knowledge to change the temporal focus of the neighborhood search based on the generation. It is empirically shown that the multitiered memetic MOEA is able to find solutions that dominate the base-line memetic algorithm. Additional experiments suggest that using local search on only non-dominated individuals improves the effectiveness and efficiency of the algorithm versus applying the local search to dominated individuals as well. This research validates the importance of using multi-objective problem (MOP) domain knowledge in order to obtain the best results for a real world solution. It also introduces a new multitiered local search procedure that is able to focus the local search on specific critical elements of the problem at different stages in the optimization process.

Analysis of a Parallel MOEA Solving the Multi-objective Quadratic Assignment Problem

The Quadratic Assignment Problem (QAP) is an NP-Complete problem [1]. The multiobjective Quadratic Assignment Problem (mQAP) is the multiobjective version of the QAP and was formalized in 2002 [2]. The QAP has had extensive research, but mQAP research is still in its infancy. The mQAP has been been used to optimize communication for formations of heterogenous unmanned aerial vehicles (UAVs) through the use of the Multi-Objective Messy Genetic Algorithm – II (MOMGA-II) [3]. This research extends that research by using a parallelized version of MOMGA-II and comparing the speedup and efficiency of the parallel results to the serial results.

Multi-objective evolutionary search performance with explicit building-block sizes for NPC problems

This research uses an explicit building block based MOEA to solve the multiobjective quadratic assignment problem. We use the multiobjective messy genetic algorithm II (MOMGA-II) to determine what role certain building blocks sizes play in filling up the Pareto front. Additionally, we investigate the role of the competitive template. The algorithm uses the competitive template by propagating it through all the building block sizes and by randomizing it for each building block size. We show that randomized competitive templates produce better results due to more exploration, and larger building block sizes are more common on the outer edges of the Pareto front because they fill more chromosome characteristics in the genotype space.

Multiobjective evolutionary algorithms for designing capacitated network centric communications

Contemporary network centric systems must provide an underlying structure for improved information communication, information awareness, sharing and collaboration between network elements. Such systems should enhance the quality of information awareness, security, improving sustainability, and mission effectiveness and efficiency. An element of network centric design is the solving of the communications or information flow problem. In this research, a multiobjective evolutionary algorithm (MOEA) is used to solve a variation of the multicommodity capacitated network design problem (MCNDP). This variation represents a hybrid communication network as found in network centric models with multiple objectives including costs, delays, robustness, vulnerability, and reliability. Nodes in such centric systems can have multiple and varying link capacities, rates and information (commodity) quantities to be delivered and received. Each commodity can have an independent prioritized bandwidth requirement as well. The nondominated sorting genetic algorithm (NSGA-II) MOEA is modified and extended to solve this generic MCNDP. Since the MCNDP is highly constrained, a novel initialization procedure and mutation method are also integrated into this MOEA. For this research, two objectives (total network cost and average number of hops) were optimized. Empirical results and analysis for 10-node and 20-node networks indicate that effective solutions can be generated efficiently. ∗ The views expressed in this article are those of the authors and do not reflect the official policy of the United States Air Force, Department of Defense, or the United States Government.