HARRIS HAWKS OPTIMIZATION DR. AHMED FOUAD ALI FACULTY OF COMPUTERS AND INFORMATICS SUEZ CANAL UNIVERSITY Outline Harris hawks optimization (HHO) (History and main idea). Social behaviors and hunting strategy. Harris hawks optimization algorithm (HHO). Diversification phase (exploration). Switch between diversification (exploration) and intensification (exploitation). Intensification phase (exploitation). Soft (smooth) besiege strategy. Hard besiege strategy. Soft (smooth) besiege strategy and progressive quick pounce. Hard besiege strategy and progressive quick pounce. Pseudo-code of Harris hawks optimization algorithm (HHO). References. Harris hawks optimization (HHO) (History and main idea) Harris hawks optimization (HHO) is a populationbased swarm intelligence algorithm which is proposed by Heidari et al. HHO mimics the hunting strategy of the Harris hawks birds. They are predators birds that are living in a group and they are hunting their prey in a smart way. Social behaviors and hunting strategy. Harris hawks are smart birds and they are living in groups. Harris's hawk has a unique foraging behavior because it attacks prey with other group members while other raptors hunt a chase alone. They are monitoring, encircling and finally attacking the prey. Social behaviors and hunting strategy (Cont.) In the morning, the individuals in the group start the hunting mission by living the reminder roosts and land on huge trees in their home kingdom. Each member of the group knows the position of the other members during the hunting process. The foraging party starts by exploring the hunting area from some group members and then perch on rather perches. Social behaviors and hunting strategy (Cont.) Harris' hawks use a surprise attack strategy to catch their prey which is a rabbit in most cases. The hawks apply different attack strategies such as long and short rapid pounce due to the high escaping capabilities of the prey in a few minutes. They transferring from one hunting strategy to another based on the escaping style of a rabbit(prey). Social behaviors and hunting strategy (Cont.) The hawks encircling the prey and attack it from different positions in order to exhaust it. Once the leader hawk (the nearest member to the prey) pounces the prey and lost it, the other members continue the chasing. Eventually, the most powerful hawk can catch the tired prey and sharing it with other group members. Harris hawks optimization algorithm (HHO) HHO has two main phases, diversification (exploration) and intensification (exploitation) which mimics the attacking strategy of Harris hawks when they hunting the prey. The attacking strategy is changed based on the circumstance of the prey. The strategies can be simulated in the HHO as follows. Diversification phase (exploration) In nature, Harris' hawks have sharp eyes that can help them to monitor and discover the prey. In HHO, the Harris' hawks represent the solutions. The best solution in each iteration represents the prey. Harris' hawks settle randomly in some places and they have two strategies to attack prey. Diversification phase (exploration) (Cont.) The first rule in Equation 1, represents the random generation of solutions. The second rule in Equation 1 represents the difference of the position of the best solution (rabbit) and the average location of the group. r3 is a random coefficient to increase the diversity of the search. r1, r2, r3, r4 and p are random numbers in (0,1). Diversification phase (exploration) (Cont.) The hawks average position can be defined as shown in Equation 2 Switch between diversification (exploration) and intensification (exploitation). The HHO algorithm can switch between diversification (exploration) and intensification (exploitation) due to the escaping energy E of the rabbit (prey). The mathematical model for the energy of prey can be defined as shown in Equation 3. Switch between diversification (exploration) and intensification (exploitation) (Cont.) The status of the prey is shown as follows. Intensification phase (exploitation) Harris' hawks execute the surprise dive by pouncing the prey. However, preys have a powerful capability to escape from a risky situation. If r is a prey's chance to escape from pouncing situations, it can be represented as follows. Soft (smooth) besiege strategy If the prey has some energy, it tries to escape from hawks by doing random jumps. However, the Harris' hawks surrounding the prey softly to exhaust it and then execute the surprise attack. This process can happen, when not successfully escaping chance r equals r ≥ 0.5 and the escaping energy of the prey E equals E ≥ 0.5. This process can be modeled as follows. Hard besiege strategy. If the prey has a little escaping energy (|E| < 0.5) and it becomes exhausted (Unsuccessfully escaping r ≥ 0.5 , the Harris hawks surround the prey and perform the surprise attack. This situation can be modeled as follows. Soft (smooth) besiege strategy and progressive quick pounce. If the prey has some energy to escape ((|E| ≥ 0.5) it can successfully escaping r < 0.5. In this case, the Harris‘ hawks apply a smooth (soft) besiege to attack the prey. The zigzag motion of the prey during the escaping process can be simulated by using a Levy flight (LF) operator. The Harris‘ hawks try to change their pouncing strategy progressively based on the tricky movements of the prey. Soft (smooth) besiege strategy and progressive quick pounce. The Harris‘ hawks can perform the soft besiege by deciding their next position as follows. The Harris‘ hawks try to adjust their movement by comparing the current pounce result and the previous one. If the result is not good, they will pounce based on the LF as follows. Soft (smooth) besiege strategy and progressive quick pounce. Based on the previous assumption of the soft besiege, Harris‘ hawks update their position as follows. Hard besiege strategy and progressive quick pounce. The Harris‘ hawks apply the hard besiege strategy when the prey has a little energy to escape (|E| < 0.5) and it has a chance to escape successfully r < 0.5. To perform this strategy, the Harris‘ hawks try to reduce the distance of their average position Xm with the prey. The overall process is shown as follows. Pseudo-code of Harris hawks optimization algorithm (HHO). Parameter setting Population initialization Population evaluation Assign the best solution Soft (smooth) besiege strategy Hard besiege strategy. Soft (smooth) besiege strategy and progressive quick pounce Hard besiege strategy and progressive quick pounce. Produce the overall best solution References A. A. Heidari, S. Mirjalili, H. Faris, I. Aljarah, M. Mafarja and H. Chen. Harris hawks optimization: Algorithm and applications. Future generation computer systems, 97, 849-872, (2019).