1 thought on “Optimized algorithm (13) Whale algorithm”
Conrad
(The following descriptions are not academic terms, for everyone to read only) Whale algorithm (Whale Algorithm) is an algorithm proposed based on the behavior of whale wrap. Whales are a group of mammals living in groups. They also cooperate with each other to drive and enclosure of prey when hunting. The whale algorithm does not propose a long time, and it is also an emerging optimization algorithm. In the whale algorithm, the position of each whale represents a feasible solution. During the hunting of the whale group, each whale has two behaviors, one is to surround the prey, and all the whales are moving towards other whales; the other is a steam bag. In each generation of swimming, whales will randomly choose these two behaviors to hunt. In the whale surrounding prey, the whale will randomly choose whether to swim towards the best whale or randomly choose a whale as its own goal and approach it. The whale algorithm, obviously, the protagonist is whale. The position of each whale in the Daries of D.
E each whale is randomly selected to surround the prey or use a steam net to drive the prey. Each whale chooses these two behaviors This rate is waiting, that is, p (surround) = p (steam bubble net) = 0.5. The whales will choose to swim towards the best whale or towards a random whale when surrounding prey. The position update formula for the whale:
The is the position of the current best whale, and each dimension of A is evenly distributed in (-a, a) The initial value of A is 2, as the iterative number decreases to 0; C is a random number distributed in (0,2) evenly. || The absolute value of the number, that is, the value of each dimension is non -negative. The position update formula for the whale:
The position of the whale selected in the current group. when will the whale choose to swim to the best individual and when will the random individual go the goal? This will be determined by the value of A . The whale chose to swim towards the best individual. Note that A is a D -dimensional vector, so when A's mold is less than 1, the whale swims towards the best individual. At the time, the whale chose to swim towards a random individual. It can be seen in the process of surrounding prey, the search mode of the whale algorithm is searching for surrounding surroundings that are closer to the best individual or nearby from random individuals. 2.2 Bubble net The whale spraying a steam bag to form a bubble net to drive the prey when hunting.
TV is the constant (no definition is found, 1 defined 1), l is a random number distributed evenly in [-1,1]. Before each action, each whale will throw a coin to decide whether to choose to surround the prey or use a bubble net to drive the prey. It's description can be seen that the process of whale algorithms is also very simple.
The adaptation function
Experiment 1: Standard whale algorithm can see from the figure that the convergence of the algorithm is still very strong, in the first place The 35th generation has completely converged. Looking at the final result, it is already a very good result, and the local search ability of the algorithm is also very strong. This confirms my statement in the previous section. The algorithm convergence is fast and lacks the ability to jump out of the best part. It from the process of the algorithm, we can see that the convergence of the algorithm is probably determined by the parameter A. Since A decreases from 2 to 0, the search range of the algorithm is getting smaller and smaller, thereby accelerating the convergence of the algorithm. This should be an optimized parameter. Now we are fixed to A to weaken the algorithm, weaken its convergence, and see if the global search and jump out of the local optimal ability have been strengthened. Experiment 2: Fixed parameter A In the image can be seen that the algorithm has almost no convergence, and the convergence speed of the algorithm is still fast. The look at the experimental results. The results than the standard whale algorithm, which can explain that parameter A affects the search accuracy of the algorithm. The effect of parameter A on the convergence of algorithm lies in the effect of A on the vector A. When A = 1.5, the mold of A has a greater probability than 1 before. At this time, the probability of choosing to swim to the optimal individual in the behavior of whales to surround the prey, so that the convergence of the algorithm is slower slower. At the same time, the global search ability of the algorithm has improved. The whale algorithm, as an emerging algorithm, I don't have too many research on it. Throughout the process of the algorithm, we can see that the standard whale algorithm and the fireflies algorithm are similar. They are all searched in the early stage of the algorithm, and local search in the later stage of the algorithm has not jumped out of the optimal operation. In the face of simple performances, the excellent performance may decline on complex issues, but because the algorithm process and structure are relatively simple, the improvement point of the algorithm is not too much. The following indicators are pure individual YY, for reference only references Mirjalili s, lewis a. The What Algorithm [J]. Advances in Software, 2016, 95: 51-67. Extract Code: B13X Catalog The above -mentioned optimization algorithm (12) Fireworks algorithm The next article optimization algorithm note (14) Water wave algorithm n 3) Whale algorithm Matlab implementation
(The following descriptions are not academic terms, for everyone to read only)
Whale algorithm (Whale Algorithm) is an algorithm proposed based on the behavior of whale wrap. Whales are a group of mammals living in groups. They also cooperate with each other to drive and enclosure of prey when hunting. The whale algorithm does not propose a long time, and it is also an emerging optimization algorithm.
In the whale algorithm, the position of each whale represents a feasible solution. During the hunting of the whale group, each whale has two behaviors, one is to surround the prey, and all the whales are moving towards other whales; the other is a steam bag. In each generation of swimming, whales will randomly choose these two behaviors to hunt. In the whale surrounding prey, the whale will randomly choose whether to swim towards the best whale or randomly choose a whale as its own goal and approach it.
The whale algorithm, obviously, the protagonist is whale.
The position of each whale in the Daries of D.
E each whale is randomly selected to surround the prey or use a steam net to drive the prey. Each whale chooses these two behaviors This rate is waiting, that is, p (surround) = p (steam bubble net) = 0.5.
The whales will choose to swim towards the best whale or towards a random whale when surrounding prey.
The position update formula for the whale:
The is the position of the current best whale, and each dimension of A is evenly distributed in (-a, a) The initial value of A is 2, as the iterative number decreases to 0; C is a random number distributed in (0,2) evenly. || The absolute value of the number, that is, the value of each dimension is non -negative.
The position update formula for the whale:
The position of the whale selected in the current group.
when will the whale choose to swim to the best individual and when will the random individual go the goal?
This will be determined by the value of A
. The whale chose to swim towards the best individual. Note that A is a D -dimensional vector, so when A's mold is less than 1, the whale swims towards the best individual.
At the time, the whale chose to swim towards a random individual.
It can be seen in the process of surrounding prey, the search mode of the whale algorithm is searching for surrounding surroundings that are closer to the best individual or nearby from random individuals.
2.2 Bubble net
The whale spraying a steam bag to form a bubble net to drive the prey when hunting.
TV is the constant (no definition is found, 1 defined 1), l is a random number distributed evenly in [-1,1].
Before each action, each whale will throw a coin to decide whether to choose to surround the prey or use a bubble net to drive the prey.
It's description can be seen that the process of whale algorithms is also very simple.
The adaptation function
Experiment 1: Standard whale algorithm
can see from the figure that the convergence of the algorithm is still very strong, in the first place The 35th generation has completely converged. Looking at the final result, it is already a very good result, and the local search ability of the algorithm is also very strong. This confirms my statement in the previous section. The algorithm convergence is fast and lacks the ability to jump out of the best part.
It from the process of the algorithm, we can see that the convergence of the algorithm is probably determined by the parameter A. Since A decreases from 2 to 0, the search range of the algorithm is getting smaller and smaller, thereby accelerating the convergence of the algorithm. This should be an optimized parameter. Now we are fixed to A to weaken the algorithm, weaken its convergence, and see if the global search and jump out of the local optimal ability have been strengthened.
Experiment 2: Fixed parameter A
In the image can be seen that the algorithm has almost no convergence, and the convergence speed of the algorithm is still fast.
The look at the experimental results.
The results than the standard whale algorithm, which can explain that parameter A affects the search accuracy of the algorithm. The effect of parameter A on the convergence of algorithm lies in the effect of A on the vector A. When A = 1.5, the mold of A has a greater probability than 1 before. At this time, the probability of choosing to swim to the optimal individual in the behavior of whales to surround the prey, so that the convergence of the algorithm is slower slower. At the same time, the global search ability of the algorithm has improved.
The whale algorithm, as an emerging algorithm, I don't have too many research on it. Throughout the process of the algorithm, we can see that the standard whale algorithm and the fireflies algorithm are similar. They are all searched in the early stage of the algorithm, and local search in the later stage of the algorithm has not jumped out of the optimal operation. In the face of simple performances, the excellent performance may decline on complex issues, but because the algorithm process and structure are relatively simple, the improvement point of the algorithm is not too much.
The following indicators are pure individual YY, for reference only
references
Mirjalili s, lewis a. The What Algorithm [J]. Advances in Software, 2016, 95: 51-67. Extract Code: B13X
Catalog
The above -mentioned optimization algorithm (12) Fireworks algorithm
The next article optimization algorithm note (14) Water wave algorithm n 3) Whale algorithm Matlab implementation