Writing Problem and Hypothesis Statements for Engineering Research(6)

8/7/2019 Writing Problem and Hypothesis Statements for Engineering Research(6)

1/12

Writing Problem and Hypothesis

Statements for Engineering Research(6)/

http://www.chineseowl.idv.tw
http://www.chineseowl.idv.tw/http://www.chineseowl.idv.tw/


2/12

Setting of work proposal: ??

Work problem:?

Quantitative specification of problem:?

Importance of problem:,?

Project need: ,?


3/12

Work objective:?

Methodology to achieve objective

:?Anticipated results:

?

Contribution to field:?


4/12

() Many machine learning and optimizationapplication-related problems are solved by using geneticalgorithms (GAs) with the multi-state property. For instance, inchess, a good player often employs various strategies based onhis opponents moves, the games progress, or the chess

clock. Therefore, an intelligent chess playing program shouldconsider the multi-state property to perform more effectively.Consider stock market investments as another example.Investors adopt various strategies according to whether themarket is up or down. This behavior implies that a decisionsupport system for investment should consider the multi-state

property. Applying various strategies to distinct states of aproblem is natural. A different solution or strategy should beemployed by varying problem solving states to achieve theglobal optimum. However, conventional methods cannot solvemulti-state problems.


5/12

() Although genetic algorithms are widely applied tomachine learning and optimization, conventional GAs have notreceived much attention. Conventional approaches can use humandesigned rule bases to represent a multi-state solution, as well as useGAs to alter the rule base. However, to our knowledge, no systematicmethod has been developed for dealing with the multi-state property in

a GA-implemented system. If the solution varies with the problem statein a multi-state problem, conventional methods neglect the multi-stateproperty and yield an inaccurate and unfeasible solution. (NOTE : Add1-3 more sentences that describe characteristics of the problem orstatistics that reflect its severity) For instance, aninvestment decision support system that adopts a single strategy,regardless of whether the market is up or down, leads to anunsuccessful investment. Although a rule basecan be used to represent the multi-state property, designing the rulebase requires many man hours. Therefore, an effectiveGAs model must be developed, capable of deriving an optimal solutionfor each state in a multi-state problem.


6/12

() An effective GAs model can be designed, capable of

deriving an optimal solution for each state in a multi-state problem.

To do so, the proposed fuzzy polyploidy, a multi-state

chromosome coding scheme, can be used to describe the solution of a

multi-state problem. An adaptive genetic structure model can then be

adopted to derive an appropriate polyploidy structure for practical

applications. The proposed model consists of three structural level

operations, including structural expansion, structural deletion, and

structural coercion, to simulate the natural random variation.

As anticipated, the proposed model can increase the accuracy of

the optimum solution derived for a particular multi-state problem.Restated, a deeper multi-state property of a problem implies a more

accurate solution achievable by the proposed model. For applications

with only a signal state (without the multi-state property), the proposed

model is comparable to conventional GAs and only requires a small

amount of extra memory and computational time.


7/12

()Moreover, the proposed GA method can enhance conventional genetic

algorithms by systematically solving multi-state problems through the

use of the polyploidy concept. Polyploidy encoding observed in nature

provides a more flexible and dynamic encoding scheme than do

conventional approaches. In practice, genetic algorithms have difficulty

in obtaining optimum solutions when the chromosome structure is too

complex because a complex structure always involves a large search

space. Importantly, the proposed simple to complex

process, accompanied by a polyploidy model, facilitates the evolution of

a complex structure that involves a large search space. (NOTE : Add

2-3 more sentences that describe more thoroughly how the proposedmethod contributes to a particular field or sector)


8/12

() As is well known, geometric parameterssignificantly affect the reliability of flip chip packaging underthermodynamic loading. Based on the basic material mechanics,a higher solder joint height implies a larger shear force that thesolder joint can bear. Moreover, several works have also

concluded that cracks occur in a solder joint when a package isunder critical thermodynamic loading. Furthermore, certainmanufacturing issues, such as packaging falling-off problems,solder bridging and misalignment are also related to the designof the pad and solder joint. However, geometricparameters of a C4 type solder joint still can not accurately

predict the reliability of flip chip packaging under thermodynamicloading. Indeed, many engineers have tried to predict thegeometric parameters of a C4 type solder joint using an energy-based method and analytical models. Although one of themsimulated the C4 type solder joint as a semi-spherical high-leadbump buried into the eutectic solder, the simulation lacks

accuracy when the external loading increases.


9/12

()Additionally, the above method disregards the effects of gravity.Moreover, the semi-spherical high-lead bump can not simulatethe high-lead bump when the solder pad sizes vary. (NOTE :Add 1-2 more sentences that describe characteristics of theproblem or statistics that reflect its severity)

For instance, if the geometric parameters can not reach anaccuracy of 5%, engineers can neither predict the fatigue life norenhance the yield of a flip chip package. Under these circumstances, engineers can not predict thecorrect geometric parameters of a C4 type solder joint when thedesign factors vary. Consequently, engineers can neither predict

the fatigue life nor enhance the yield of a flip chip package. Therefore, an analytical geometry method must bedeveloped, capable of accurately predicting the geometricparameters of practical C4 type solder joint in flip chiptechnology after a reflow process.


10/12

() An analytical geometry method can bedevised, capable of accurately predicting the geometricparameters of practical C4 type solder joint in flip chiptechnology after a reflow process. To doso, the differential equation of curvature can be derived

from the well-known Laplace-Young equation. Additionally,the contact angle between the solder and substrate and thesurface tension of the solder can be measured. The freesurface energy of the substrate and the interfacial tensionbetween the solder and the substrate can then be obtained

from Youngs equation. Next, the differential equation ofcurvature can be numerically solved to obtain thegeometric parameters. Moreover, the geometric data of thesolder joint can be used as an input in the finite elementmodel to analyze the stress/strain distribution, thermalfatigue life and reliability of the electronic packages.


11/12


12/12

Further details can be found at

http://www.chineseowl.idv.tw
http://www.chineseowl.idv.tw/http://www.chineseowl.idv.tw/

Writing Problem and Hypothesis Statements for Engineering Research(6)

Documents