A OPEN PIT MINING AÇIK OCAK MADENCİLİĞİ

23 rd
producing the new vector x
i , G 1
for the next
generation.
The main advantage of jDE algorithm is
that, the user does not need to guess the good
values for F and CR at the beginning of the
algorithm. These parameters are implicitly
adapted inside the algorithm.
In this paper, jDE algorithm is used to
optimize the blending of ore blocks in open
pit mining.
4.2 Fitness Function of Blend
Optimization
The mathematical model of the blending
optimization is given in equations 2-5. In
order to solve the blending optimization
problem using a genetic algorithm, the
concepts introduced by Olivetti (2003) and
Huy et al. (2004), is practiced here, to define
the fitness function.
D is the set of optimization parameters (set
of decision variables) which is called an
individual and in the case of blending
optimization, D is equal to the number of ore
block. In genetic algorithms, fitness function
gives a score for each individual
(chromosome) which defines the probability
of that individual to be chosen for the next
generation.
Considering equation 2-5 and assuming
that there is only one impurity in the ore
blocks, the fitness function will be calculated
as follow
Doing this, the constraints of the model in
equations 3 and 4 will be imported to the
objective function. Using the fitness function
described in equation 9, the score of each
individual can be calculated in the jDE
algorithm.
5 APPLICATION IN GOL-E-GOHAR
IRON ORE MINE
Gol-e-Gohar iron ore mine number 2 is
selected to check the applicability of the pitblend
algorithm to determine the pit limits of
this mine. Gol-e-Gohar iron ore mine
number 2 is located in south-east of Iran
(figure 2). The block model of the deposit is
constructed using 42 drill holes. Small block
dimension are decided to be 10*10*15
meters. The number of block on each axis of
the large block is 210*140*20 blocks. Each
block contains information about its sulfur,
phosphor and Fe content. The shape of the
deposit is depicted in figure 3. The blocks
are divided into two main ore types, one
which is low in sulfur and the other which is
high in sulfur content. The block model
contains 8259 blocks of both ore types. The
block model designed for this mine has got
up to 588’000 blocks.
fvalue POP T C C (9)
In which
i T g g im im
T X
bx
bB
b
C
g
Max 0,
gb Gmin
x
b
bB
C
im
Max 0,
gb immax
x
b
bB
T
,
g
, and
im
are the degree of importance
(or weight) of T, C
g
, and C
im
in the fitness
function respectively. POP
i
, i 1,...,
NP is a
candidate solution of the problem.
Figure 2. Location of Gol-e-Gohar iron ore
mines
Prior to applying the pit-blend optimizing
procedure, bounding technique was applied
to discard the unnecessary blocks of the
block model. The bounded block model
125