Interactive Product Design Selection With an Implicit Value Function

2005 ◽  
Vol 127 (3) ◽  
pp. 367-377 ◽  
Author(s):  
K. Maddulapalli ◽  
S. Azarm ◽  
A. Boyars

We present a new method to aid a decision maker (DM) in selecting the “most preferred” from a set of design alternatives. The method is deterministic and assumes that the DM’s preferences reflect an implicit value function that is quasi-concave. The method is interactive, with the DM stating preferences in the form of attribute tradeoffs at a series of trial designs, each a specific design under consideration. The method is iterative and uses the gradient of the value function obtained from the preferences of the DM to eliminate lower value designs at each trial design. We present an approach for finding a new trial design at each iteration. We provide an example, the design selection for a cordless electric drill, to demonstrate the method. We provide results showing that (within the limit of our experimentation) our method needs only a few iterations to find the most preferred design alternative. Finally we extend our deterministic selection method to account for uncertainty in the attributes when the probability distributions governing the uncertainty are known.

Author(s):  
K. Maddulapalli ◽  
S. Azarm ◽  
A. Boyars

We present an automated method to aid a Decision Maker (DM) in selecting the ‘most preferred’ from a set of design alternatives. The method assumes that the DM’s preferences reflect an implicit value function that is quasi-concave. The method is iterative, using three approaches in sequence to eliminate lower-value alternatives at each trial design. The method is interactive, with the DM stating preferences in the form of attribute tradeoffs at each trial design. We present an approach for finding a new trial design at each iteration. We provide an example, the design selection for a cordless electric drill, to demonstrate the method.


2005 ◽  
Vol 128 (5) ◽  
pp. 1027-1037 ◽  
Author(s):  
A. K. Maddulapalli ◽  
S. Azarm

An important aspect of engineering product design selection is the inevitable presence of variability in the selection process. There are mainly two types of variability: variability in the preferences of the decision maker (DM) and variability in attribute levels of the design alternatives. We address both kinds of variability in this paper. We first present a method for selection with preference variability alone. Our method is interactive and iterative and assumes only that the preferences of the DM reflect an implicit value function that is differentiable, quasi-concave and non-decreasing with respect to attributes. The DM states his/her preferences with a range (due to the variability) for marginal rate of substitution (MRS) between attributes at a series of trial designs. The method uses the range of MRS preferences to eliminate “dominated designs” and then to find a set of “potentially optimal designs.” We present a payload design selection example to demonstrate and verify our method. Finally, we extend our method for selection with preference variability to the case where the attribute levels of design alternatives also have variability. We assume that the variability in attribute levels can be quantified with a range of attribute levels.


2011 ◽  
Author(s):  
Anouk Festjens ◽  
Siegfried Dewitte ◽  
Enrico Diecidue ◽  
Sabrina Bruyneel

2021 ◽  
Vol 14 (3) ◽  
pp. 130
Author(s):  
Jonas Al-Hadad ◽  
Zbigniew Palmowski

The main objective of this paper is to present an algorithm of pricing perpetual American put options with asset-dependent discounting. The value function of such an instrument can be described as VAPutω(s)=supτ∈TEs[e−∫0τω(Sw)dw(K−Sτ)+], where T is a family of stopping times, ω is a discount function and E is an expectation taken with respect to a martingale measure. Moreover, we assume that the asset price process St is a geometric Lévy process with negative exponential jumps, i.e., St=seζt+σBt−∑i=1NtYi. The asset-dependent discounting is reflected in the ω function, so this approach is a generalisation of the classic case when ω is constant. It turns out that under certain conditions on the ω function, the value function VAPutω(s) is convex and can be represented in a closed form. We provide an option pricing algorithm in this scenario and we present exact calculations for the particular choices of ω such that VAPutω(s) takes a simplified form.


Author(s):  
Humoud Alsabah ◽  
Agostino Capponi ◽  
Octavio Ruiz Lacedelli ◽  
Matt Stern

Abstract We introduce a reinforcement learning framework for retail robo-advising. The robo-advisor does not know the investor’s risk preference but learns it over time by observing her portfolio choices in different market environments. We develop an exploration–exploitation algorithm that trades off costly solicitations of portfolio choices by the investor with autonomous trading decisions based on stale estimates of investor’s risk aversion. We show that the approximate value function constructed by the algorithm converges to the value function of an omniscient robo-advisor over a number of periods that is polynomial in the state and action space. By correcting for the investor’s mistakes, the robo-advisor may outperform a stand-alone investor, regardless of the investor’s opportunity cost for making portfolio decisions.


Author(s):  
Vijitashwa Pandey ◽  
Deborah Thurston

Design for disassembly and reuse focuses on developing methods to minimize difficulty in disassembly for maintenance or reuse. These methods can gain substantially if the relationship between component attributes (material mix, ease of disassembly etc.) and their likelihood of reuse or disposal is understood. For products already in the marketplace, a feedback approach that evaluates willingness of manufacturers or customers (decision makers) to reuse a component can reveal how attributes of a component affect reuse decisions. This paper introduces some metrics and combines them with ones proposed in literature into a measure that captures the overall value of a decision made by the decision makers. The premise is that the decision makers would choose a decision that has the maximum value. Four decisions are considered regarding a component’s fate after recovery ranging from direct reuse to disposal. A method on the lines of discrete choice theory is utilized that uses maximum likelihood estimates to determine the parameters that define the value function. The maximum likelihood method can take inputs from actual decisions made by the decision makers to assess the value function. This function can be used to determine the likelihood that the component takes a certain path (one of the four decisions), taking as input its attributes, which can facilitate long range planning and also help determine ways reuse decisions can be influenced.


Sign in / Sign up

Export Citation Format

Share Document