Clark School Home UMD
Home

CONTACT INFORMATION:

Professor Shapour Azarm
2155 Martin Hall
Dept of Mechanical Engineering
University of Maryland
College Park, MD 20742

Voice: +1-301-405-5250
Email: azarm@umd.edu



Research Interests

 
Funding

Organizations that have funded the research of Dr. Azarm include:

NSF
NASA
ONR
PI (UAE)
Northrop Grumman
Black & Decker
Maryland Industrial Partnerships
and others

Our research interests are on exploring models and methods in:

  1. Design for Market Systems, including design of single products, product lines, and bundled products,
  2. Multi-Criteria Decision Making, including multi-objective design optimization and multi-attribute decision analysis for finding the “best” design solution(s), and
  3. Multidisciplinary Design Optimization, including optimization-based methods for designing systems that can be decomposed into multiple subsystems.

Design for Market Systems (DMS)

Our goal in DMS has been to devise models and methods that can facilitate product design selection for single products, product lines, and bundled products. This research involves an evolving framework that integrates engineering design, marketing, service and other aspects of supply chain. A few of the questions that we have been investigating include:

  • How should the design of a single product/product line be selected under uncertainty?
  • How can products be designed to ensure a retail channel acceptance while maximizing profit in the face of customer demand and market competition?
  • What is an optimal way to devise bundling strategies for multi-category products?
  • How should online customer reviews be considered for product design selection with supply chain considerations?

figure 1
An overview of our research in the area of DMS (see, e.g., Williams, N., P. K. Kannan, and S. Azarm, 2011,
“Retail Channel Structure Impact on Strategic Engineering Product Design,” Management Science, 57(5), pp. 897-914)


Multiple Criteria Decision Making (MCDM)

Our goal in the MCDM area is to investigate models and methods that involve finding the “best” system alternative from a set of alternatives that is most attractive over multiple criteria. MCDM is divided into two sub-areas:

  1. Multi-Objective Design Optimization (MODO), and
  2. Multi-Attribute Decision Making (MADM).

Our research in MODO explores models and methods in design optimization with approximation and uncertainty considerations. For example, in approximation-assisted MODO, our approach can be used for design of engineering systems that require the use of computationally expensive analyzers (like CFD, FEM). Our methods in multi-objective robust optimization can obtain "robust solutions" that are both optimum and relatively insensitive to uncertainty. Our research in MADM is on decision support systems that can be used for selection according to several criteria and based on the decision maker’s preferences and risk attitude.

figure 2
Schematic of an approach for approximation assisted optimization (see, e.g., Li, G., S. Azarm, A. Farhang-Mehr, and
A. Diaz, 2006, “Approximation of Multi-Response Engineering Simulations: A Dependent Meta-Modeling Approach,”
Structural and Multidisciplinary Optimization, 31, pp. 260-269)


figure 3
An approximation assisted multi-objective robust optimization approach (see, e.g., Hu, W., M. Li, S. Azarm, and
A. Almansoori, 2011, “Multi-Objective Robust Optimization with Interval Uncertainty Using Online
Approximation and Constraint Cuts,” Journal of Mechanical Design, 130(6), pp. 061002-1 to 061002-9)


figure 4
MADM: An approach for making a selection (see, e.g., Maddulapalli, K., S. Azarm, and A. Boyars, 2007,
“Sensitivity Analysis for Product Design Selection with an Implicit Value Function,”
European Journal of Operational Research, 180, pp. 1245-1259)


Multidisciplinary Design Optimization (MDO)

The primary focus of our research in MDO has been on optimization techniques and models that are applicable to a broad class of problems, beyond those in which the existing techniques are applicable. More specifically, our MDO research has been focused on investigating coordination strategies and handling couplings in multi-objective MDO problems that can facilitate design of complex engineering systems that are decomposable to multiple subsystems including those with a combination of discrete and continuous variables and with uncertainty considerations.

figure 5
Schematic of a multi-objective MDO framework. The grey rectangle is for input uncertainty and the
yellow rectangle is for an acceptable output variation (see, e.g., Li, M., and S. Azarm, 2008,
“Multiobjective collaborative Robust Optimization (McRO) with Interval Uncertainty and Interdisciplinary
Uncertainty Propagation,” Journal of Mechanical Design, 130(8), pp. 081402-1 to 081402-12)



MIMOSA has the ability to analyze the effects of uncertainty on system and subsystem performance for coupled multi-output multidisciplinary problems (see, Li, M., J. Hamel, and S. Azarm, 2010, “Optimal Uncertainty Reduction for Multi-Disciplinary
Multi-Output Systems Using Sensitivity Analysis,” Structural and Multidisciplinary Optimization, 40, pp. 77-96)


Applications:
Numerous engineering examples have been used in our research to explore and demonstrate the applicability of techniques developed by our group. These include: consumer electronics (e.g., smart phones, tablet computer); heat exchangers; power tools; power electronics; data centers; oil refineries; unmanned vehicles (walking robots, UAVs, UUVs); and many others.

figure 6
Effects of different market structures on design of cordless angle grinders (see, e.g., Williams, N.,
P. K. Kannan, and S. Azarm, 2011, “Retail Channel Structure Impact on Strategic Engineering Product Design,”
Management Science, 57(5), pp. 897-914)


figure 7
An application of multi-objective robust optimization to an example in reactor-distillation chemical process (see, e.g., Hu, W.,
M. Li, S. Azarm, and A. Almansoori, 2011, “Multi-Objective Robust Optimization with Interval Uncertainty Using Online
Approximation and Constraint Cuts,” Journal of Mechanical Design, 130(6), pp. 061002-1 to 061002-9)


figure 8
Optimizing data center cabinets (see, e.g., Li, G., M. Li, S. Azarm, J. Rambo, and Y. Joshi, 2007,
“Optimizing Thermal Design of Data Center Cabinets with a New Multi-Objective Genetic Algorithm,”
Distributed and Parallel Databases, 21(2-3), pp. 167-192)

Top of page