Information Technology Reference
In-Depth Information
Fesanghary et al [80] compared IHS and GA for the optimal design of a shell and tube
heat exchanger, which included both annualised capital and operating costs.
The optimal operation of a system of cogeneration (combined heat and power)
plants was studied in [81]. The aim was to determine the heat and power generation
rates at each plant to minimise the total cost, subject to constraints on the total heat
and power demand, and the feasible operating region for each plant.
3.8 Medical Studies
Some digital hearing aids can classify their acoustic environment and adapt the
sounds transmitted to the wearer accordingly. Amor et al [82] used HS to determine
the best subset from a set of 74 standard features of the acoustic environment to be
used in the sound classifier.
High dose rate brachytherapy uses the temporary placement of tiny radiation sources
into tissue using catheters for the treatment of cancer. Panchal [83] reported on the HS
optimisation of the dwell time, which is the duration that the radiation source is left in
the vicinity of the affected tissue. Dong et al [84] studied the detection of abnormalities
in biological tissue from digital images using an adaptive-parameter HS algorithm.
3.9 Other Applications
Liu and Feng [85] used HS for system identification. They estimated ten discrete-
valued parameters of a Controlled Auto-Regressive Moving Average (CARMA)
model for oil well heat wash data. Two applications in process control have been reported:
nonlinear model predictive control for set-point tracking using HAA [86], and synchroni-
zation of discrete-time chaotic systems using another modified HS method [87].
Further applications include the solution of Sudoku puzzles [88], musical composition
[89], timetabling and room allocation for university courses [25], optimization of a milling
process [90], and the selection of land parcels for ecological conservation (a MCSP) [17].
4 Developments in the HS Method
This section first presents a classification for the modifications made to the original HS
method, and then it shows which modified algorithms use which particular innovations.
Lastly, developments in the mathematical analysis of the HS method are reported.
4.1 Classification of Modifications to the Original HS
The original HS algorithm was presented in Chapter 1, and we briefly recap the pa-
rameters and steps in the algorithm here.
Harmony Memory (HM) is the principal HS data structure. It is a matrix that stores
in its rows a selection of the current best harmonies (solution vectors):
⎡
1
⎤
x
HM
#
⎢
⎥
(5)
HMS
×
n
=
∈
ℜ
⎢
⎥
⎢
⎥
HMS
x
⎣
⎦
Search WWH ::
Custom Search