Cloud computing task scheduling method based on improved African bald eagle algorithm

Abstract

The invention discloses a task scheduling method, and particularly relates to a cloud computing task scheduling method based on an improved African bald eagle algorithm, which comprises the following steps of: S1, initializing algorithm parameters, and enabling an AVOA population to correspond to a cloud computing task scheduling scheme; s2, judging whether the algorithm is ended or not, and if not, executing the step S3; s3, traversing the population, evaluating the fitness value of each vulture, and finding out optimal and sub-optimal vulture positions; s4, updating parameters, and selecting an exploration strategy or a development strategy according to the value of the vulture starvation degree F; and S5, introducing an adaptive opposite learning strategy to update the position of the vulture, and returning to the step S2. According to the cloud computing task scheduling method based on the improved African bald eagle algorithm, a better task allocation scheme can be quickly found in a cloud system, and the task execution efficiency and the resource utilization rate are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a cloud computing task scheduling method based on an improved African vulture algorithm. Background Art

[0002] In order to reduce the computing burden of terminal computers, the processing power of the "cloud" can be improved to minimize the computing power occupied by the terminal. Cloud computing can reduce the cost of various hardware devices (such as memory, processor, graphics card) during the development process of enterprises. However, when enterprises use cloud computing technology in large quantities, energy consumption will increase and the waiting time of users may also be extended. How to reduce energy consumption and delay is a hot topic in the development of cloud computing. Task scheduling based on intelligent algorithms is a very important scheduling method for cloud computing. Task completion time is one of the important optimization goals of task scheduling. Reducing task completion time can improve the operating efficiency of the system and user satisfaction.

[0003] The present invention performs resource modeling and task modeling on cloud computing task scheduling, and designs a suitable fitness function to evaluate the quality of task allocation schemes. The present invention adopts AVOA (African vulture algorithm) to schedule cloud tasks. At the beginning of the algorithm, the vulture population is initialized by the Tent reverse learning strategy to obtain an initial solution with uniform distribution. An adaptive adversarial learning strategy is introduced into the African vulture algorithm to effectively reduce the task completion time and better avoid falling into the local optimal solution. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention provides a cloud computing task scheduling method based on an improved African vulture algorithm to solve the current technical problems.

[0005] The technical solution adopted by the present invention to solve its technical problem is: The present invention provides a cloud computing task scheduling method based on an improved African vulture algorithm, comprising the following steps: Step S1: Initialize algorithm parameters and correspond the AVOA population to the cloud computing task scheduling scheme; Step S2: Determine whether to end the algorithm, if not, execute step S3; Step S3: traverse the population, evaluate the fitness value of each vulture, and find the best and second best vulture positions; Step S4: Update parameters and select exploration strategy or development strategy according to the value of vulture hunger degree |F|; Step S5: Introduce the adaptive adversarial learning strategy to update the vulture position and return to step S2.

[0006] Preferably, the step S1 includes the following sub-steps: Step S11: Establish a cloud task scheduling model; Step S12: Setting the algorithm parameters of the African vulture algorithm, including the maximum number of iterations , population size N, P 1 , P 2 , P 3 , ub upper bound and lb lower bound; Step S13: Initialize the vulture population through the Tent reverse learning strategy to generate a certain number of vultures, each vulture represents a possible task scheduling solution.

[0007] Preferably, the step S13 includes the following sub-steps: Step S131: tent chaos initialization; ; ; in, is the i-th mapping function value, i is the number of mappings, for The population individuals obtained by inverse mapping to the individual search space, ub and lb are the upper and lower bounds of the search space; Step S132: generating a reverse population; ; in, Represents the solution generated by the Tent reverse learning strategy corresponding to the i-th individual in the population; Step S133: Merge selection; merge the population generated by the Tent chaotic map with the reverse population to form a new population set, calculate the fitness value of each individual in the new population, sort them according to the fitness value, and select the first N individuals with better fitness values ​​as the final initial population.

[0008] Preferably, step S3 includes the following sub-steps: Step S31: Setting the vulture fitness function and the time fitness function; Step S32: Select the vultures with the best position and the second best position.

[0009] Preferably, step S4 includes the following sub-steps: Step S41: setting hunger rate function F; Step S42: When |F|≥1, it means that the vulture has enough physical strength to find food, and step S43 is executed, and the vulture enters the exploration strategy; when |F|<1, step S44 is executed; Step S43: exploration strategy; Step S44: Develop strategy.

[0010] Preferably, step S5 includes the following sub-steps: Step S51: introducing an adaptive adversarial learning strategy to prevent the vulture individuals from falling into a local optimal solution; Step S52: Calculate and compare the fitness values ​​of the current position and the positions of the opposing solutions, and select the solution with the smaller fitness value as the new position of the vulture according to the greedy strategy; Step S53: Return to step S2.

[0011] The beneficial effects of the present invention are as follows: the present invention is a cloud computing task scheduling method based on the improved African vulture algorithm, which can quickly find a better task allocation solution in the cloud system, improve task execution efficiency and resource utilization, and introduce the Tent reverse learning strategy when the population of the African vulture algorithm is initialized to generate an initial population with uniform distribution and strong diversity, thereby improving the search efficiency and global optimization ability of the algorithm and improving the task scheduling efficiency. In the iterative process of the vulture, an adaptive adversarial learning strategy is introduced to enhance the search ability of the algorithm, which is conducive to jumping out of the local optimal solution and further improving the performance of cloud computing task scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which: Figure 1 is a step diagram of a cloud computing task scheduling method based on an improved African vulture algorithm according to an embodiment of the present invention; Figure 2 It is a step diagram of the initialization of the Tent reverse learning strategy population of the cloud computing task scheduling method based on the improved African vulture algorithm in an embodiment of the present invention; Figure 3 It is a step diagram of adaptive adversarial learning of a cloud computing task scheduling method based on an improved African vulture algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] like Figure 1 The figure shows a step diagram of a cloud computing task scheduling method based on an improved African vulture algorithm of the present invention, and the specific implementation steps are as follows: S11. Establishing a cloud task scheduling model can be understood as m subtasks executing tasks on n virtual machines, where (m>n). The set of m subtasks can be expressed as T={t 1 ,t 2 ,t 3 ,....,t m}, t i (i=1,2,3,...,m) represents the ith subtask; the set of n virtual machines can be represented as VM={vm 1 ,vm 2 ,vm 3 ,...,vm n}, vm j (j=1,2,3,...,n) represents the jth virtual machine. Then the task-virtual machine allocation relationship is represented by the m×n matrix TVM, as shown in formula (1): (1); Among them, tvm ij ∈{0,1}, when tvm ij When it indicates 1, the virtual machine vm j The subtask t is assigned to i ; When it is 0, the virtual machine vm j No subtask t is assigned to i For example: tvm 13 =1, it means that task 1 is assigned to virtual 3. ij represents the execution time of the task on the virtual machine, and represents the subtask t i In the virtual machine vm j Execution time on (2); (3); askTime ij is the execution time of a single subtask, length i is the length of subtask i, in mips j is the processing capacity of virtual machine j, S is the total number of cloud tasks on the current virtual machine. The total cloud task completion time is the time when the last task of the cloud system is completed.

[0015] (4).

[0016] Step S12: Initialize the algorithm parameters of the African vulture algorithm, including the maximum number of iterations , population size N, P 1 , P 2 , P 3 , ub upper bound and lb lower bound.

[0017] Step S13. Initialize the vulture population through the Tent reverse learning strategy to generate a certain number of vultures, each vulture represents a possible task scheduling scheme. In the AVOA (African Vulture Algorithm) algorithm, population initialization is very important. Most algorithms choose to use random initialization when initializing the population, but the randomly generated population is uneven in the entire space. In order to generate an initial population with uniform distribution and strong diversity, the Tent reverse learning strategy is used to generate an initial population with uniform distribution.

[0018] Further, step S2: determine whether the iteration end condition is met, if so, output the optimal solution, which is the task scheduling solution, if not, execute step S3.

[0019] Furthermore, in step S3, the fitness value of each vulture is evaluated, and the Time function in step S11 is used as the fitness function.

[0020] Step S31. Set the vulture fitness function. The time fitness function is as shown in formula (5): (5); The present invention aims to minimize the total completion time of cloud tasks in the cloud system, so Time is used as the total completion time of cloud tasks as the fitness function, calculated by formula (4). The smaller the value of the fitness function fit is, the better the performance of the task scheduling algorithm is.

[0021] S32. Select the vulture with the best and second best position according to formula (6).

[0022] (6); (7); Where: R(i) is one of the best vultures generated, Best 1 (i) represents the best individual of the first group of vultures in the current iteration round, Best 2 (i) represents the best vulture in the second group of vultures in the current iteration round, P i Calculated by formula (7), L 1 and L 2 are two random numbers in the range [0,1], the sum of the two numbers is 1, fit i represents the fitness value of the first and second groups of vultures, and n represents the total number of vulture individuals in the first and second groups. The step S4: updating parameters, calculating the hunger rate of each vulture, and selecting an exploration strategy or a development strategy according to the value of the vulture's hunger degree |F|.

[0023] Step S41. Set the hunger rate function. The vulture will enter the exploration or development stage according to the change of the hunger rate, and use the position update formula of different stages to update its own position. The hunger rate function is shown in formula (9): (8); (9); In the formula, t is a tuning parameter used to prevent falling into a local optimal solution; h is a random variable whose value range is limited to [-2,2]; Iterator i represents the current iteration round; maxiterator represents the maximum number of iterations; w is the phase adjustment decision parameter for adjusting the decision allocation between the exploration phase and the development phase. It is constant. When the w value increases, the algorithm tends to enter the exploration phase in the optimization phase. On the contrary, when the w value decreases, the probability of entering the exploration phase is reduced. In formula (9), F represents the hunger degree of the vulture; rand 1 is a random number between 0 and 1; z is a random number between -1 and 1 that changes in each iteration; Step S42. The vulture makes a decision based on its hunger level and chooses a strategy to update its position. When |F|≥1, it means that the vulture has enough physical strength to find food, and step S43 is executed, and the vulture enters the exploration strategy. When |F|<1, step S44 is executed. The vulture does not have the physical strength to support long-distance flight to find food, so they prefer to forage in nearby areas. At this time, the vulture enters the development stage.

[0024] Step S431. When It indicates that the leading vulture updates the position of the other following vultures by equation (10). (10); (11); D(i) represents the distance between the vulture and the current optimal vulture, P(i) is the current position of the individual in the population. R(i) is calculated by formula (6), F is the vulture hunger rate calculated by formula (9), and X is a random number in the range of [0,2], indicating the position where the vulture moves randomly, and is obtained using the formula X=2*rand.

[0025] Step S432. When That is, when the leader vulture conducts independent exploration, the remaining vultures update their positions in the vicinity of the leader vulture by randomly jumping within the upper and lower bounds of its search space. 1 The value of , the AVOA algorithm can simulate the foraging behavior of vultures more effectively.

[0026] (12); P1 represents the vulture decision parameter, which is set to 0.6; p(i+1) is the vulture position vector in the next iteration; rand p1 is a random number between 0 and 1; rand 2 and rand 3 Random numbers between 0 and 1, rand 3 is a coefficient used to increase randomness. 3 Taking a number close to 1, it will distribute solutions in a similar pattern. Otherwise, the strategy will search in a random search space. lb and ub represent the upper and lower bounds of the variable respectively.

[0027] Step S44. The vulture enters the early stage of the development strategy, which adopts two different food protection mechanisms.

[0028] Step S441. When When , the food competition strategy is used to update the position, as shown in formula (13).

[0029] (13); (14); D(i) is calculated by formula (11), F is the vulture hunger rate calculated by formula (9), rand 4 is a random number between (0,1), d(t) represents the distance between the optimal vulture and the current vulture, P(i) is the current position of the individual in the population, and R(i) is calculated by formula (6).

[0030] Step S442. When The rotating flight strategy is adopted to update the position according to formula (15).

[0031] (15); (16); (17); rand 5 and rand 6 is a random number between (0,1), P i Calculated by formula (7), P(i) is the current position of the individual in the population, S 1 and S 2 The vulture's rotating flight is calculated by equations (16) and (17).

[0032] Step S443. When |F| is less than 0.5, the AVOA algorithm enters the late stage of development strategy. At this time, the algorithm simulates the behavior of vultures and executes two different aggregation behaviors and food siege strategies. The selection of these strategies is based on the decision parameter p 3 Adjust if P 3 ≧rand p3 , the food source accumulation strategy is used to update the position, as shown in formula (18), which represents the vulture's behavior of gathering at the food source.

[0033] (18); (19); (20); Step S444. Calculated by formulas (19) and (20), and are the optimal solution and suboptimal solution in formula (6), P(i) is the current position of the individual in the population, and F is the vulture hunger rate calculated by formula (9). When the vultures are in a fierce competition for food, they use a fierce food competition strategy to update their positions, as shown in formula (22), which indicates the siege behavior of vultures in competing for food.

[0034] (twenty one); (twenty two); (twenty three); (twenty four); in, Calculated by formula (24), d(t) by formula (14), R(i) by formula (6), p 3 Set to 0.6, rand p3 is a random number in [0,1], u and v are random numbers between [0,1], β is the default fixed number 1.5, Γ is the gamma function, and levy(d) is the levy-flight function, which is mainly used to improve the computational efficiency of LAVOA.

[0035] Furthermore, in step S5, an adaptive adversarial learning strategy is introduced to prevent the vulture individuals from falling into the local optimal solution and balance the exploration and development of the algorithm.

[0036] Step S51: Introduce an adaptive adversarial learning strategy to prevent individual vultures from falling into local optimal solutions, balance the exploration and development of the algorithm, and generate an adversarial solution for each vulture. In adversarial learning, an adaptive factor is introduced to adjust the generation strategy of adversarial solutions.

[0037] (25); (26); P(i) is the current position of the individual in the population, is the adaptive factor, which gradually decreases with the increase of the number of iterations t, where t is the current number of iterations, and t 0 is a constant that controls the shape of the function, and k is a constant that controls the rate at which the function changes.

[0038] Step S52: Calculate and compare the fitness values ​​of the current position and the positions of the opposing solutions, and select the solution with the smaller fitness value as the new position of the vulture according to the greedy strategy.

[0039] Step S53. Return to step S2.

[0040] like Figure 2 The figure shows the initialization steps of the population of the Tent reverse learning strategy of the present invention, and the specific implementation steps are as follows: Step S131. Tent chaos initialization.

[0041] (27); (28); is the i-th mapping function value, i is the number of mappings, for The population individuals obtained by inverse mapping to the individual search space, and are the upper and lower bounds of the search space.

[0042] Step S132: Generate a reverse population.

[0043] (29); Represents the solution generated by the Tent reverse learning strategy corresponding to the i-th individual in the population.

[0044] Step S133. Merge selection: merge the population generated by the Tent chaotic map with the reverse population to form a new population set, calculate the fitness value of each individual in the new population, sort them according to the fitness value, and select the first N individuals with better fitness values ​​as the final initial population.

[0045] like Figure 3 The figure shows the steps of adaptive adversarial learning of the present invention, and the specific implementation steps are as follows: Step S51: Introduce an adaptive adversarial learning strategy to prevent individual vultures from falling into local optimal solutions, balance the exploration and development of the algorithm, and generate an adversarial solution for each vulture. In adversarial learning, an adaptive factor is introduced to adjust the generation strategy of adversarial solutions.

[0046] (30); (31); P(i) is the current position, is the adaptive factor, which gradually decreases with the increase of the number of iterations t, where t is the current number of iterations, and t 0 is a constant that controls the shape of the function, and k is a constant that controls the rate at which the function changes.

[0047] Step S52: Calculate and compare the fitness values ​​of the current position and the positions of the opposing solutions, and select the solution with the smaller fitness value as the new position of the vulture according to the greedy strategy.

[0048] (32); P(i) new is the starting point of the next iteration selected according to the greedy strategy, P opppie is the position of the opposing solution generated by the vulture, and P(i) is the current position of the individual in the population.

[0049] Step S53. Return to step S2.

[0050] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0051] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A cloud computing task scheduling method based on an improved African vulture algorithm, characterized in that: The following steps are involved: Step S1: Initialize algorithm parameters and correspond the AVOA population to the cloud computing task scheduling scheme; Step S2: Determine whether to end the algorithm, if not, execute step S3; Step S3: traverse the population, evaluate the fitness value of each vulture, and find the best and second best vulture positions; Step S4: Update parameters and select exploration strategy or development strategy according to the value of vulture hunger degree |F|; Step S5: Introduce the adaptive adversarial learning strategy to update the vulture position and return to step S2.

2. The cloud computing task scheduling method based on the improved African vulture algorithm according to claim 1 is characterized in that: The step S1 includes the following sub-steps: Step S11: Establish a cloud task scheduling model; Step S12: Setting the algorithm parameters of the African vulture algorithm, including the maximum number of iterations , population size N, P1, P2, P3, ub upper bound and lb lower bound; Step S13: Initialize the vulture population through the Tent reverse learning strategy to generate a certain number of vultures, each vulture represents a possible task scheduling solution.

3. The cloud computing task scheduling method based on the improved African vulture algorithm according to claim 2 is characterized in that: The step S13 includes the following sub-steps: Step S131: tent chaos initialization; ； ； in, is the i-th mapping function value, i is the number of mappings, for The population individuals obtained by inverse mapping to the individual search space, ub and lb are the upper and lower bounds of the search space; Step S132: generating a reverse population; ； in, Represents the solution generated by the Tent reverse learning strategy corresponding to the i-th individual in the population; Step S133: Merge selection; merge the population generated by the Tent chaotic map with the reverse population to form a new population set, calculate the fitness value of each individual in the new population, sort them according to the fitness value, and select the first N individuals with better fitness values ​​as the final initial population.

4. The cloud computing task scheduling method based on the improved African vulture algorithm according to claim 1 is characterized in that: The step S3 includes the following sub-steps: Step S31: Setting the vulture fitness function and the time fitness function; Step S32: Select the vultures with the best position and the second best position.

5. The cloud computing task scheduling method based on the improved African vulture algorithm according to claim 1 is characterized in that: The step S4 includes the following sub-steps: Step S41: setting hunger rate function F; Step S42: When |F|≥1, it means that the vulture has enough physical strength to find food, and step S43 is executed, and the vulture enters the exploration strategy; when |F|<1, step S44 is executed; Step S43: exploration strategy; Step S44: Develop strategy.

6. The improved cloud computing task scheduling method based on the African vulture algorithm according to claim 1 is characterized in that: The step S5 includes the following sub-steps: Step S51: introducing an adaptive adversarial learning strategy to prevent the vulture individuals from falling into a local optimal solution; Step S52: Calculate and compare the fitness values ​​of the current position and the positions of the opposing solutions, and select the solution with the smaller fitness value as the new position of the vulture according to the greedy strategy; Step S53: Return to step S2.

Images