Gabriella Dellino

Kriging metamodel management in the design optimization of a CNG injection system

This paper deals with the use of Kriging metamodels in multi-objective engineering design optimization. The metamodel management issue to find the tradeoff between accuracy and efficiency is addressed. A comparative analysis of different strategies is conducted for a case study devoted to the design of a component of the injection system for Compressed Natural Gas (CNG) engines. The computational results are reported and analyzed for a performance assessment conducted with a data envelopment analysis approach.

Robust simulation-optimization using metamodels

Optimization of simulated systems is the goal of many methods, but most methods assume known environments. In this paper we present a methodology that does account for uncertain environments. Our methodology uses Taguchis view of the uncertain world, but replaces his statistical techniques by either Response Surface Methodology or Kriging metamodeling. We illustrate the resulting methodology through the well-known Economic Order Quantity (EOQ) model.

A decomposition approach for the combined master surgical schedule and surgical case assignment problems

This research aims at supporting hospital management in making prompt Operating Room (OR) planning decisions, when either unpredicted events occur or alternative scenarios or configurations need to be rapidly evaluated. We design and test a planning tool enabling managers to efficiently analyse several alternatives to the current OR planning and scheduling. To this aim, we propose a decomposition approach. More specifically, we first focus on determining the Master Surgical Schedule (MSS) on a weekly basis, by assigning the different surgical disciplines to the available sessions. Next, we allocate surgeries to each session, focusing on elective patients only. Patients are selected from the waiting lists according to several parameters, including surgery duration, waiting time and priority class of the operations. We performed computational experiments to compare the performance of our decomposition approach with an (exact) integrated approach. The case study selected for our simulations is based on the characteristics of the operating theatre (OT) of a medium-size public Italian hospital. Scalability of the method is tested for different OT sizes. A pilot example is also proposed to highlight the usefulness of our approach for decision support. The proposed decomposition approach finds satisfactory solutions with significant savings in computation time.

Appliance operation scheduling for electricity consumption optimization

This paper concerns the problem of optimally scheduling a set of appliances at the end-user premises. The users energy fee varies over time, and moreover, in the context of smart grids, the user may receive a reward from an energy aggregator if he/she reduces consumption during certain time intervals. In a household, the problem is to decide when to schedule the operation of the appliances, in order to meet a number of goals, namely overall costs, climatic comfort level and timeliness. We devise a model accounting for a typical household user, and present computational results showing that it can be efficiently solved in real-life instances.

Optimization models for consumer flexibility aggregation in smart grids: The ADDRESS approach

This paper addresses the problem of optimal management of consumer flexibility in an electric distribution system. Aggregation of a number of consumers clustered according to appropriate criteria, is one of the most promising approaches for modifying the daily load profile at nodes of an electric distribution network. Modifying the daily load profile is recognized as one of the strongest needs both for safe and efficient operation of the network. The paper proposes an optimization approach allowing the aggregator, i.e., the operator which manages the aggregated consumers, to gather flexibility and generate bids for the energy market, with the aim of maximizing its revenue. It is shown that this problem can be solved through mixed integer linear programming. Numerical simulation results are provided for validating the proposed approach.

Energy production forecasting in a PV plant using transfer function models

This paper deals with the issue of forecasting energy production of a Photo-Voltaic (PV) plant, needed by the Distribution System Operator (DSO) for grid planning. As the energy production of a PV plant is strongly dependent on the environmental conditions, the DSO has difficulties to manage an electrical system with stochastic generation. This implies the need to have a reliable forecasting of the irradiance level for the next day in order to setup the whole distribution network. To this aim, this paper proposes the use of transfer function models. The assessment of quality and accuracy of the proposed method has been conducted on a set of scenarios based on real data.

Enhanced Evolutionary Algorithms for Multidisciplinary Design Optimization: A Control Engineering Perspective

Summary. This chapter deals with the application of hybrid evolutionary methods to design optimization issues in which approximation techniques and model management strategies can be used to guide the decision making process in a multidisciplinary context. An enhanced evolutionary algorithmic scheme devoted to design optimization is proposed, and its use in real applications is illustrated in the framework of the multidisciplinary design optimization (MDO). At this aim, a case study is discussed. It relies to the field of automotive engineering in which the design optimization of a system is carried out considering simultaneously both mechanical and control requirements. The studied system is the regulator of the injection pressure of an innovative common rail system for compressed natural gas (CNG) engines, whose engineering design optimization includes several practical and numerical difficulties. To tackle such a situation, a multiobjective optimization formulation of the problem is proposed. The adopted optimization strategy pursues the Pareto optimality on the basis of fitness functions that capture domain specific design aspects as well as static and dynamic objectives. The computational experiments show the ability of the proposed method for finding a satisfactory set of efficient solutions.

Multidisciplinary design optimization of a pressure controller for CNG injection systems

In this work, the multidisciplinary design optimization (MDO) methodology is applied to a case arising in the automotive engineering in which the design optimization of mechanical and control features of a system are simultaneously carried out with an evolutionary algorithm based method. The system under study is the regulator of the injection pressure of an innovative Common Rail system for Compressed Natural Gas (CNG) automotive engines, whose engineering design includes several practical and numerical difficulties. To tackle such a situation, this paper proposes a constrained multi-objective optimization method, that pursues the Pareto-optimality on the basis of fitness functions that capture domain specific design aspects as well as static and dynamic objectives. The proposed scheme provides ways to incorporate the designers specific knowledge, from interactive actions to simulation based analysis or surrogate-assisted evolution. The computational experiments show the ability of the method for finding a relevant and satisfactory set of efficient solutions.

Metamodel-Based Robust Simulation-Optimization: An Overview

Optimization of simulated systems is the goal of many techniques, but most of them assume known environments. Recently, “robust” methodologies accounting for uncertain environments have been developed. Robust optimization tackles problems affected by uncertainty, providing solutions that are in some sense insensitive to perturbations in the model parameters. Several alternative methods have been proposed for achieving robustness in simulation-based optimization problems, adopting different experimental designs and/or metamodeling techniques. This chapter reviews the current state of the art on robust optimization approaches based on simulated systems. First, we summarize robust Mathematical Programming. Then we discuss Taguchi’s approach introduced in the 1970s. Finally, we consider methods to tackle robustness using metamodels, and Kriging in particular. The proposed methodology uses Taguchi’s view of the uncertain world, but replaces his statistical techniques by Kriging. We illustrate the resulting methodology through basic inventory models.

Parametric and distribution-free bootstrapping in robust simulation-optimization

Most methods in simulation-optimization assume known environments, whereas this research accounts for uncertain environments combining Taguchis world view with either regression or Kriging (also called Gaussian Process) metamodels (emulators, response surfaces, surrogates). These metamodels are combined with Non-Linear Mathematical Programming (NLMP) to find robust solutions. Varying the constraint values in this NLMP gives an estimated Pareto frontier. To account for the variability of this estimated Pareto frontier, this contribution considers different bootstrap methods to obtain confidence regions for a given solution. This methodology is illustrated through some case studies selected from the literature.