A Case Study of Stochastic Optimization in Health Policy: Problem Formulation and Preliminary Results

We use Bayesian decision theory to address a variable selection problem arising in attempts to indirectly measure the quality of hospital care, by comparing observed mortality rates to expected values based on patient sickness at admission. Our method weighs data collection costs against predictive accuracy to find an optimal subset of the available admission sickness variables. The approach involves maximizing expected utility across possible subsets, using Monte Carlo methods based on random division of the available data into N modeling and validation splits to approximate the expectation. After exploring the geometry of the solution space, we compare a variety of stochastic optimization methods –- including genetic algorithms (GA), simulated annealing (SA), tabu search (TS), threshold acceptance (TA), and messy simulated annealing (MSA) –- on their performance in finding good subsets of variables, and we clarify the role of N in the optimization. Preliminary results indicate that TS is somewhat better than TA and SA in this problem, with MSA and GA well behind the other three methods. Sensitivity analysis reveals broad stability of our conclusions.     

Solving a dynamic cell formation problem using metaheuristics

In this paper, solving a cell formation (CF) problem in dynamic condition is going to be discussed by using some traditional metaheuristic methods such as genetic algorithm (GA), simulated annealing (SA) and tabu search (TS). Most of previous researches were done under the static condition. Due to the fact that CF is a NP-hard problem, then solving the model using classical optimization methods needs a long computational time. In this research, a nonlinear integer model of CF is first given and then solved by GA, SA and TS. Then, the results are compared with the optimal solution and the efficiency of the proposed algorithms is discussed.     

Parallelizing simulated annealing algorithms based on high-performance computer

We implemented five conversions of simulated annealing (SA) algorithm from sequential-to-parallel forms on high-performance computers and applied them to a set of standard function optimization problems in order to test their performances. According to the experimental results, we eventually found that the traditional approach to parallelizing simulated annealing, namely, parallelizing moves in sequential SA, difficultly handled very difficult problem instances. Divide-and-conquer decomposition strategy used in a search space sometimes might find the global optimum function value, but it frequently resulted in great time cost if the random search space was considerably expanded. The most effective way we found in identifying the global optimum solution is to introduce genetic algorithm (GA) and build a highly hybrid GA+SA algorithm. In this approach, GA has been applied to each cooling temperature stage. Additionally, the performance analyses of the best algorithm among the five implemented algorithms have been done on the IBM Beowulf PCs Cluster and some comparisons have been made with some recent global optimization algorithms in terms of the number of functional evaluations needed to obtain a global minimum, success rate and solution quality.     

Utilizing expected improvement and generalized data envelopment analysis in multi-objective genetic algorithms

Meta-heuristic methods such as genetic algorithms (GA) and particle swarm optimization (PSO) have been extended to multi-objective optimization problems, and have been observed to be useful for finding good approximate Pareto optimal solutions. In order to improve the convergence and the diversity in the search of solutions using meta-heuristic methods, this paper suggests a new method to make offspring by utilizing the expected improvement (EI) and generalized data envelopment analysis (GDEA). In addition, the effectiveness of the proposed method will be investigated through several numerical examples in comparison with the conventional multi-objective GA and PSO methods.     

Evaluating Performance of Honey Bee Mating Optimization

The honey bee mating optimization (HBMO) algorithm is presented and tested with various test functions, and its performance is compared with the genetic algorithm (GA). It is shown that the HBMO algorithm can overcome the weaknesses of the GA. The HBMO converges faster than the GA. Even when the HMBO starts from a more improper initial condition than the GA, it can reach a better solution in a smaller number of function evaluations. Furthermore, in some cases, the GA was not able to reach the global minimum.     

FRACTOP: A Geometric Partitioning Metaheuristic for Global Optimization

We propose a new metaheuristic, FRACTOP, for global optimization. FRACTOP is based on the geometric partitioning of the feasible region so that search metaheuristics such as Simulated Annealing (SA), or Genetic Algorithms (GA) which are activated in smaller subregions, have increased reliability in locating the global optimum. FRACTOP is able to incorporate any search heuristic devised for global optimization. The main contribution of FRACTOP is that it provides an intelligent guidance (through fuzzy measures) in locating the subregion containing the global optimum solution for the search heuristics imbedded in it. By executing the search in nonoverlapping subregions, FRACTOP eliminates the repetitive visits of the search heuristics to the same local area and furthermore, it becomes amenable for parallel processing. As FRACTOP conducts the search deeper into smaller subregions, many unpromising subregions are discarded from the feasible region. Thus, the initial feasible region gains a fractal structure with many space gaps which economizes on computation time. Computational experiments with FRACTOP indicate that the metaheuristic improves significantly the results obtained by random search (RS), SA and GA.     

BP-GA混合优化策略在人力资源战略规划中的应用

采用混合优化策略训练神经网络,进而实现地区人力资源数据的时间序列预测.神经网络,尤其是应用反向传播(back propagation,简称BP)算法训练的神经网络,被广泛应用于预测中.但是BP神经网络训练速度慢、容易陷入局部极值.遗传算法(genetic algorithm,简称GA)具有很好的全局寻优性.因而提出将BP和GA结合起来的混合优化策略训练神经网络,来实现人力资源数据预测.与BP算法相比,数值计算结果表明预测精度高、速度快,为地区人力资源数据的时间序列预测研究提供了一条新的途径.     

Genetic local search for multi-objective combinatorial optimization

The paper presents a new genetic local search (GLS) algorithm for multi-objective combinatorial optimization (MOCO). The goal of the algorithm is to generate in a short time a set of approximately efficient solutions that will allow the decision maker to choose a good compromise solution. In each iteration, the algorithm draws at random a utility function and constructs a temporary population composed of a number of best solutions among the prior generated solutions. Then, a pair of solutions selected at random from the temporary population is recombined. Local search procedure is applied to each offspring. Results of the presented experiment indicate that the algorithm outperforms other multi-objective methods based on GLS and a Pareto ranking-based multi-objective genetic algorithm (GA) on travelling salesperson problem (TSP).     

The use of genetic algorithms to solve the economic lot size scheduling problem

The purpose of this paper is to investigate the use genetic algorithms (GAs) for solving the Economic Lot Size Scheduling Problem (ELSP). The ELSP is formulated using the Basic Period (BP) approach which results in a problem having one continuous decision variable and a number of integer decision variables equal to the number of products being produced. This formulation is ideally suited for using GAs. The GA is tested on Bomberger's classical problem. The resulting solutions were better than those obtained using an iterative dynamic programming (DP) approach. The total cost of GA solutions to the problem with utilization up to 65% were within 3.4% of the lower bound. The GA also performed well for higher utilization yielding solutions within 13.87% of the lower bound for utilization up to 86%. The GA was tested on a 30-item problem and good solutions were obtained. The results of the GA under different binary representations, crossover methods, and initialization methods are compared to identify the best settings. The results indicate that for this particular problem, binary representation works better than Gray coding, 2-point crossover is best, and an infeasible starting population is better than feasible.     

Integer Programming Duality in Multiple Objective Programming

The weighted sums approach for linear and convex multiple criteria optimization is well studied. The weights determine a linear function of the criteria approximating a decision makers overall utility. Any efficient solution may be found in this way. This is not the case for multiple criteria integer programming. However, in this case one may apply the more general e-constraint approach, resulting in particular single-criteria integer programming problems to generate efficient solutions. We show how this approach implies a more general, composite utility function of the criteria yielding a unified treatment of multiple criteria optimization with and without integrality constraints. Moreover, any efficient solution can be found using appropriate composite functions. The functions may be generated by the classical solution methods such as cutting plane and branch and bound algorithms.     

Embedding evolutionary strategy in ordinal optimization for hard optimization problems

This work proposes a method for embedding evolutionary strategy (ES) in ordinal optimization (OO), abbreviated as ESOO, for solving real-time hard optimization problems with time-consuming evaluation of the objective function and a huge discrete solution space. Firstly, an approximate model that is based on a radial basis function (RBF) network is utilized to evaluate approximately the objective value of a solution. Secondly, ES associated with the approximate model is applied to generate a representative subset from a huge discrete solution space. Finally, the optimal computing budget allocation (OCBA) technique is adopted to select the best solution in the representative subset as the obtained “good enough” solution. The proposed method is applied to a hotel booking limits (HBL) problem, which is formulated as a stochastic combinatorial optimization problem with a huge discrete solution space. The good enough booking limits, obtained by the proposed method, have promising solution quality, and the computational efficiency of the method makes it suitable for real-time applications. To demonstrate the computational efficiency of the proposed method and the quality of the obtained solution, it is compared with two competing methods – the canonical ES and the genetic algorithm (GA). Test results demonstrate that the proposed approach greatly outperforms the canonical ES and GA.     

A genetic algorithm for robotic assembly line balancing

Flexibility and automation in assembly lines can be achieved by the use of robots. The robotic assembly line balancing (RALB) problem is defined for robotic assembly line, where different robots may be assigned to the assembly tasks, and each robot needs different assembly times to perform a given task, because of its capabilities and specialization. The solution to the RALB problem includes an attempt for optimal assignment of robots to line stations and a balanced distribution of work between different stations. It aims at maximizing the production rate of the line. A genetic algorithm (GA) is used to find a solution to this problem. Two different procedures for adapting the GA to the RALB problem, by assigning robots with different capabilities to workstations are introduced: a recursive assignment procedure and a consecutive assignment procedure. The results of the GA are improved by a local optimization (hill climbing) work-piece exchange procedure. Tests conducted on a set of randomly generated problems, show that the Consecutive Assignment procedure achieves, in general, better solution quality (measured by average cycle time). Further tests are conducted to determine the best combination of parameters for the GA procedure. Comparison of the GA algorithm results with a truncated Branch and Bound algorithm for the RALB problem, demonstrates that the GA gives consistently better results.     

On a Feasible–Infeasible Two-Population (FI-2Pop) genetic algorithm for constrained optimization: Distance tracing and no free lunch

We explore data-driven methods for gaining insight into the dynamics of a two-population genetic algorithm (GA), which has been effective in tests on constrained optimization problems. We track and compare one population of feasible solutions and another population of infeasible solutions. Feasible solutions are selected and bred to improve their objective function values. Infeasible solutions are selected and bred to reduce their constraint violations. Interbreeding between populations is completely indirect, that is, only through their offspring that happen to migrate to the other population. We introduce an empirical measure of distance, and apply it between individuals and between population centroids to monitor the progress of evolution. We find that the centroids of the two populations approach each other and stabilize. This is a valuable characterization of convergence. We find the infeasible population influences, and sometimes dominates, the genetic material of the optimum solution. Since the infeasible population is not evaluated by the objective function, it is free to explore boundary regions, where the optimum is likely to be found. Roughly speaking, the No Free Lunch theorems for optimization show that all blackbox algorithms (such as Genetic Algorithms) have the same average performance over the set of all problems. As such, our algorithm would, on average, be no better than random search or any other blackbox search method. However, we provide two general theorems that give conditions that render null the No Free Lunch results for the constrained optimization problem class we study. The approach taken here thereby escapes the No Free Lunch implications, per se.     

Experiments with LAGRASP heuristic for set k-covering

The set k-covering problem (SC _k P) is a variant of the classical set covering problem, in which each object is required to be covered at least k times. We describe a hybrid Lagrangean heuristic, named LAGRASP, which combines subgradient optimization and GRASP with path-relinking to solve the SC _k P. Computational experiments carried out on 135 test instances show experimentally that by properly tuning the parameters of LAGRASP, it is possible to obtain a good trade-off between solution quality and running times. Furthermore, LAGRASP makes better use of the dual information provided by subgradient optimization and is able to discover better solutions and to escape from locally optimal solutions even after the stabilization of the lower bounds, whereas other strategies fail to find new improving solutions.     

A branch-and-bound-based local search method for the flow shop problem

It is well-known that exact branch and bound methods can only solve small or moderately sized ????-hard combinatorial optimization problems. In this paper, we address the issue of embedding an approximate branch and bound algorithm into a local search framework. The resulting heuristic has been applied to the problem of finding a minimum makespan in the permutation flow shop problem. Computational experiments carried out on a large set of benchmark problems show that the proposed method consistently yields optimal or near-optimal solutions for instances with up to 200 jobs and 10 machines. In particular, for 19 instances, the heuristic produces solutions that outperform the best known ones.     

NeuroGenetic approach for combinatorial optimization: an exploratory analysis

Given the NP-Hard nature of many optimization problems, it is often impractical to obtain optimal solutions to large-scale problems in reasonable computing time. For this reason, heuristic and metaheuristic search approaches are used to obtain good solutions fast. However, these techniques often struggle to develop a good balance between local and global search. In this paper we propose a hybrid metaheuristic approach which we call the NeuroGenetic approach to search for good solutions for these large scale optimization problems by at least partially overcoming this challenge. The proposed NeuroGenetic approach combines the Augmented Neural Network (AugNN) and the Genetic Algorithm (GA) search approaches by interleaving the two. We chose these two approaches to hybridize, as they offer complementary advantages and disadvantages; GAs are very good at searching globally, while AugNNs are more proficient at searching locally. The proposed hybrid strategy capitalizes on the strong points of each approach while avoiding their shortcomings. In the paper we discuss the issues associated with the feasibility of hybridizing these two approaches and propose an interleaving algorithm. We also provide empirical evidence demonstrating the effectiveness of the proposed approach.     

A Hybrid Metaheuristic for the Quadratic Assignment Problem

The quadratic assignment problem (QAP) is known to be NP-hard. We propose a hybrid metaheuristic called ANGEL to solve QAP. ANGEL combines the ant colony optimization (ACO), the genetic algorithm (GA) and a local search method (LS). There are two major phases in ANGEL, namely ACO phase and GA phase. Instead of starting from a population that consists of randomly generated chromosomes, GA has an initial population constructed by ACO in order to provide a good start. Pheromone acts as a feedback mechanism from GA phase to ACO phase. When GA phase reaches the termination criterion, control is transferred back to ACO phase. Then ACO utilizes pheromone updated by GA phase to explore solution space and produces a promising population for the next run of GA phase. The local search method is applied to improve the solutions obtained by ACO and GA. We also propose a new concept called the eugenic strategy intended to guide the genetic algorithm to evolve toward a better direction. We report the results of a comprehensive testing of ANGEL in solving QAP. Over a hundred instances of QAP benchmarks were tested and the results show that ANGEL is able to obtain the optimal solution with a high success rate of 90%. This work was supported in part by the National Science Council, R.O.C., under Contract NSC 91-2213-E-005-017.     

一种类微粒群算法及其在混流装配线调度问题中的应用

混流装配生产线的调度问题是准时化生产系统中最重要的问题之一.基于微粒群算法的原理,提出了一种类微粒群算法——PPSO(Pseudo Particle Swarm Optimization),可应用于解决准时化生产方式下的混流装配线调度问题.数值试验表明,采用PPSO方法比采用目标追随法、遗传算法和模拟退火算法的求解质量有很大提高.     

Design method using hybrid of line-type and circular-type routes for transit network system optimization

Transit network design is a very important problem. In particular, it has a great influence on passenger satisfaction with the whole transit network system. The present research proposes a simulated annealing (SA) method for optimizing a transit network design. In the algorithm, the strategy to search for neighborhood solutions provides the chance to find the best hybrid of line-type and circular-type routes. The proposed SA method is also compared with other methods. The results show that the proposed SA model is a good alternative for transit network design, particularly as it provides the scope to design hybrids of line-type and circular-type routes.     

Economic optimization in a fixed sequence of unreliable inspections

Given a fixed sequence of unreliable inspection operations with known costs and inspection error probabilities of two types (classifying good items as defective and vice versa), we develop a model for selecting the set of inspections that should be activated in order to minimize expected total costs (inspection and penalties). We present an efficient branch and bound algorithm for finding the optimal solution, and two variations of a greedy heuristic that can be applied jointly to provide very good solutions at a O(n²) computational complexity. The conclusions are backed by a factorial experiment that included 1440 problem instances.