Haobin Li

MO-COMPASS: a fast convergent search algorithm for multi-objective discrete optimization via simulation

Discrete Optimization via Simulation (DOvS) has drawn considerable attention from both simulation researchers and industry practitioners, due to its wide application and significant effects. In fact, DOvS usually implies the need to solve large-scale problems, making the efficiency a key factor when designing search algorithms. In this research work, MO-COMPASS (Multi-Objective Convergent Optimization via Most-Promising-Area Stochastic Search) is developed, as an extension of the single-objective COMPASS, for solving DOvS problems with two or more objectives by taking into consideration the Pareto optimality and the probability of correct selection. The algorithm is proven to be locally convergent, and numerical experiments have been carried out to show its ability to achieve high convergence rate.

Multi-objective COMPASS for discrete optimization via simulation

Due to its wide application in many industries, discrete optimization via simulation (DOvS) has recently attracted more research interests. As industry systems become more complex, advanced search algorithms for DOvS are desired with higher expectation towards efficiency. In this research work, we combine the ideas of single-objective Convergent Optimization via Most-Promising-Area Stochastic Search (COMPASS) with the concept of Pareto optimality to propose multi-objective (MO) MO-COMPASS for solving DOvS problems with two or more objectives. Numerical experiments are illustrated to show its ability to achieve high efficiency.

A study on multi-objective particle swarm optimization with weighted scalarizing functions

In literature, multi-objective particle swarm optimization (PSO) algorithms are shown to have great potential in solving simulation optimization problems with real number decision variables and objectives. This paper develops a multi-objective PSO algorithm based on weighted scalarization (MPSOws) in which objectives are scalarized by different sets of weights at individual particles while evaluation results are shared among the swarm. Various scalarizing functions, such as simple weighted aggregation (SWA), weighted compromise programming (WCP), and penalized boundary intersection (PBI) can be applied in the algorithm. To improve the diversity and uniformity of the Pareto set, a hybrid external archiving technique consisting of both KNN and ε-dominance methods is proposed. Numerical experiments on noise-free problems are conducted to show that MPSOws outperforms the benchmark algorithm and WCP is the most preferable strategy for the scalarization. In addition, simulation allocation rules (SARs) can be further applied with MPSOws when evaluation error is considered.

Multi-objective optimization for a hospital inpatient flow process via discrete event simulation

This paper describes a Discrete Event Simulation (DES) model for a hypothetical inpatient flow process of a large acute-care hospital. The implementation of the Multi-Objectives Convergent Optimization via Most-Promising-Area Stochastic Search (MO-COMPASS) approach in this DES model for the identification of promising Pareto optimal solutions is also discussed. The MO-COMPASS algorithm implemented within the DES modelling paradigm demonstrates how the Multi-Objective Discrete Optimization via Simulation (MDOvS) framework can be applied to identify process improvement opportunities for a hypothetical inpatient boarding processes of a large acute care hospital.

The object-oriented discrete event simulation modeling: a case study on aircraft spare part management

Object-Oriented DES (O2DES) is an effort to implement the object oriented paradigm in the scope of ease the development of discrete event simulation models in both education as well as industrial settings. In particular, O2DES offers several functionalities which support the integration of the tool with optimization techniques, thus making it easier to the students to understand the concept of simulation-optimization. It also supports the application of different variance reduction techniques such as budget allocation and time dilation. In order to do so, the provided toolkit exploits the C# language and the .NET Framework and it guarantees the efficient generation of DES models, as well as the effectiveness of the developed models in being integrated with sampling solutions. We propose a case study related to the aircraft spare part management problem to show case the main functionalities of the proposed tool.

Multi-objective multi-fidelity optimization with ordinal transformation and optimal sampling

In simulation-optimization, the accurate evaluation of candidate solutions can be obtained by running a high-fidelity model, which is fully featured but time-consuming. Less expensive and lower fidelity models can be particularly useful in simulation-optimization settings. However, the procedure has to account for the inaccuracy of the low fidelity model. Xu et al. (2015) proposed the MO2TOS, a Multi-fidelity Optimization (MO) algorithm, which introduces the concept of ordinal transformation (OT) and uses optimal sampling (OS) to exploit models of multiple fidelities for efficient optimization. In this paper, we propose MO-MO2TOS for the multi-objective case using the concepts of non-dominated sorting and crowding distance to perform OT and OS in this setting. Numerical experiments show the satisfactory performance of the procedure while analyzing the behavior of MO-MO2TOS under different consistency scenarios of the low-fidelity model. This analysis provides insights on future studies in this area.

Constrained optimizaton for hospital bed allocation via discrete event simulation with nested partitions

This paper aims to further motivate the use of simulation of complex systems in optimizing healthcare operations under uncertainty. One argument to use optimization only such as mathematical programming instead of simulation optimization in making decisions is the ability of the former to account for constraints and to consider a large number of alternatives. However, current state-of-the art of simulation optimization has opened the possibilities of using both simulation and optimization in the case of multiple performance measures. We consider the case of hospital bed allocation and give an example on how a stochastically constrained optimization via simulation can be applied. Nested Partitions are used for the search algorithm and combined with OCBA-CO, an efficient simulation budget allocation, as simulation is time-consuming.

Optimization via gradient oriented polar random search

Search algorithms are often used for optimization problems where its mathematical formulation is difficult to be analyzed, e.g., simulation optimization. In literature, search algorithms are either driven by gradient or based on random sampling within specified neighborhood, but both methods have limitation as gradient search can be easily trapped at a local optimum and random sampling loses efficiency by not utilizing local information such as gradient direction that might be available. A combination of the two is believed to overcome both disadvantages. However, the main difficulty is how to incorporate and control randomness in a direction instead of a point. Thus, this paper makes use of a polar coordinate representation in any high dimension to randomly generate directions where the concentration can be explicitly controlled, based on which a brand new Gradient Oriented Polar Random Search (GO-POLARS) is designed and proved to satisfy the conditions for strong local convergence.

Optimal design of master-worker architecture for parallelized simulation optimization

This study formulates and solves the design problem for a master-worker architecture dedicated to the implementation of a parallelized simulation optimization algorithm. Such a formulation does not assume any specific characteristic of the optimization problem being solved, but the way the algorithm is parallelized. In particular, we refer to the master-worker paradigm, where the master makes sampling decisions while the workers receive solutions to evaluate. We identify two metrics to be optimized: the throughput of the workers in terms of the number of evaluations per time unit, and the lack of synchronization between the master and the workers. We identify several design parameters: number of workers (n), the buffer size for each worker and for the master and the sample size m, i.e., the number of solutions used by the master for sampling decisions at each iteration. Numerical experiments show optimal designs over randomly generated simulation optimization algorithm instances.

Capacity planning for mega container terminals with multi-objective and multi-fidelity simulation optimization

ABSTRACT Container terminals play a significant role as representative logistics facilities for contemporary trades by handling outbound, inbound, and transshipment containers to and from the sea (shipping liners) and the hinterland (consignees). Capacity planning is a fundamental decision process when constructing, expanding, or renovating a container terminal to meet demand, and the outcome of this planning is typically represented in terms of configurations of resources (e.g., the numbers of quay cranes, yard cranes, and vehicles), which enables the container flows to satisfy a high service level for vessels (e.g., berth-on-arrivals). This study presents a decision-making process that optimizes the capacity planning of large-scale container terminals. Advanced simulation-based optimization algorithms, such as Multi-Objective Multi-Fidelity Optimization with Ordinal Transformation and Optimal Sampling (MO-MO2TOS), Multi-Objective Optimal Computing Budget Allocation (MOCBA), and Multi-Objective Convergent Optimization via Most-Promising-Area Stochastic Search (MO-COMPASS), were employed to formulate and optimally solve the large-scale multi-objective problem with multi-fidelity simulation models. Various simulation results are compared with one another in terms of the capacities over different resource configurations to understand the effect of various parameter settings on optimal capacity across the algorithms.