A hybrid optimization technique coupling an evolutionary and a local search algorithm

Evolutionary algorithms are robust and powerful global optimization techniques for solving large-scale problems that have many local optima. However, they require high CPU times, and they are very poor in terms of convergence performance. On the other hand, local search algorithms can converge in a few iterations but lack a global perspective. The combination of global and local search procedures should offer the advantages of both optimization methods while offsetting their disadvantages. This paper proposes a new hybrid optimization technique that merges a genetic algorithm with a local search strategy based on the interior point method. The efficiency of this hybrid approach is demonstrated by solving a constrained multi-objective mathematical test-case.     

Variable selection for multivariate calibration using a genetic algorithm: prediction of additive concentrations in polymer films from Fourier transform-infrared spectral data

Variable selection using a genetic algorithm is combined with partial least squares (PLS) for the prediction of additive concentrations in polymer films using Fourier transform-infrared (FT-IR) spectral data. An approach using an iterative application of the genetic algorithm is proposed. This approach allows for all variables to be considered and at the same time minimizes the risk of overfitting. We demonstrate that the variables selected by the genetic algorithm are consistent with expert knowledge. This very exciting result is a convincing application that the algorithm can select correct variables in an automated fashion.     

Using genetic algorithms to characterize ferrofluid topographies in externally applied magnetic fields

Both ferrofluidics and genetic algorithms are relatively new fields. Due to complex physical interactions, ferrofluidic topographies and assemblies have only been solved using finite time step, Lattice Boltzmann, and finite-element methods in very simple magnetic field configurations. In this paper, we show that it is possible (and highly advantageous) to employ genetic algorithms to solve for the fluid topographies, which can be extended to include more complex magnetic fields.     

Stationary Probability Distribution of Colored-noise Saturation Model for a Single-mode Dye Laser

It is applied the interpolation procedure to calculate the stationary probability distribution of colored-gain-noise model of a single-mode dye laser which operates above threshold with correlation time τ covering a very wide rang. By use of Stochastic Runge-Kutta Algorithm, it also has carried out numerical simulations of the steady-state properties. Comparing the results of the interpolation procedure and the unified colored-noise approximation with simulation results, the agreement between the results of the interpolation procedure and simulation results is much better than that of the unified colored-noise approximation when correlation time τ covers range from moderate to large.     

Periodic schedules for linear precedence constraints

We consider the computation of periodic cyclic schedules for linear precedence constraints graphs: a linear precedence constraint is defined between two tasks and induces an infinite set of usual precedence constraints between their executions such that the difference of iterations is a linear function. The objective function is the minimization of the maximal period of a task.We recall first that this problem may be modelled using linear programming. A polynomial algorithm is then developed to solve it for a particular class of linear precedence graphs called unitary graphs. We also show that a periodic schedule may not exist for unitary graphs. In the general case, a decomposition of the linear precedence graph into unitary components is computed and we assume that a periodic schedule exists for each of these components. Lower bounds on the periods are exhibited and we show that an optimal periodic schedule may not achieve them. The notion of quasi-periodic schedule is then introduced and we prove that this new class of schedules always reaches these bounds.     

A genetic algorithm for the two-stage supply chain distribution problem associated with a fixed charge

This paper considers a two-stage distribution problem of a supply chain that is associated with a fixed charge. Two kinds of cost are involved in this problem: a continuous cost that linearly increases with the amount transported between a source and a destination, and secondly, a fixed charge, that incurs whenever there exists a transportation of a non-zero quantity between a source and a destination. The objective criterion is the minimisation of the total cost of distribution. A genetic algorithm (GA) that belongs to evolutionary search heuristics is proposed and illustrated. The proposed methodology is evaluated for its solution quality by comparing it with the approximate and lower bound solutions. Thus, the comparison reveals that the GA generates better solution than the approximation method and is capable of providing solution either equal or closer to the lower bound solution of the problem.     

Genetic algorithms and cellular automata in aquifer management

A groundwater management problem is presented involving pumping cost minimization with both well discharges and well locations as decision variables. A grid of candidate well locations is set up and optimal arrangements of wells are sought within this discrete space. A genetic algorithm approach is presented with the following particular features: (a) A suitable scaling is applied to the objective function in order to alleviate its regionally flat behavior. (b) No penalty functions are involved in constraint handling. Instead, the feasible region is transformed into a rectangular domain. The transformation introduced is proved to be bijective. (c) A binary representation of well configurations is presented and compared to a combinatorial one. The binary representation necessitates the introduction of specially designed genetic operators. Besides purely genetic algorithms, the concept of cellular automaton is introduced as the basis of an alternative formulation of the optimization problem. The lattice of the cellular automaton provides the discrete set of candidate well positions. The well configuration is represented by a group of agents occupying an equal number of lattice sites. The agents change positions as dictated by the structure of the automaton and, also, by an associated genetic algorithm, which directs the evolution of the whole scheme toward an optimal configuration. An improved performance of this approach is noted and discussed in comparison to the purely genetic algorithm schemes of the present work. A simulated annealing approach is also applied to the same problem for comparison purposes. Finally, a new and more efficient hybrid annealing–genetic approach is introduced and discussed.     

A Hybrid Genetic Algorithm with Boltzmann Convergence Properties

Stochastic global search algorithms such as genetic algorithms are used to attack difficult combinatorial optimization problems. However, genetic algorithms suffer from the lack of a convergence proof. This means that it is difficult to establish reliable algorithm braking criteria without extensive a priori knowledge of the solution space. The hybrid genetic algorithm presented here combines a genetic algorithm with simulated annealing in order to overcome the algorithm convergence problem. The genetic algorithm runs inside the simulated annealing algorithm and provides convergence via a Boltzmann cooling process. The hybrid algorithm was used successfully to solve a classical 30-city traveling salesman problem; it consistently outperformed both a conventional genetic algorithm and a conventional simulated annealing algorithm. This work was supported by the University of Colorado at Colorado Springs.     

Variable neighborhood search for minimum cost berth allocation

The berth allocation problem is to allocate space along the quayside to incoming ships at a container terminal in order to minimize some objective function. We consider minimization of total costs for waiting and handling as well as earliness or tardiness of completion, for all ships. We assume ships can arrive at any given time, i.e., before or after the berths become available. The resulting problem, which subsumes several previous ones, is expressed as a linear mixed 0–1 program. As it turns out to be too time-consuming for exact solution of instances of realistic size, a Variable Neighborhood Search (VNS) heuristic is proposed, and compared with Multi-Start (MS), a Genetic Search algorithm (GA) and a Memetic Search algorithm (MA). VNS provides optimal solutions for all instances solved to optimality in a previous paper of the first two authors and outperforms MS, MA and GA on large instances.     

An algorithmic study of the Maximum Flow problem: A comparative statistical analysis

In this paper, we show the results of an experimental study about the most important algorithms proposed to solve the Maximum Flow problem. The appropriate statistical analysis not only allows us to justify comparisons between the different procedures but also to obtain classifications of their practical efficiency. Furthermore, an empirical experiment allows us to identify the influence of several parameters that are not included in a theoretical study.