Search

There are two problems in the development of swarm-based metaheuristic. First, there are not any metaheuristic is able to solve all problems superiorly. Second, the cross-dimension mapping between the entity and its reference during performing the directed search is rare to find. Based on these problems, this work introduces a new metaphorfree swarm-based metaheuristic called the triple-shake algorithm (TSA). As its name suggests, TSA consists of three directed searches. The reference in the first search is the balance mixture between the finest entity and the member of permutation set. The reference in the second search is the balance mixture between the finest entity and a randomly chosen entity. The reference in the third search is the finest entity only. But the cross-dimension mapping is performed in this third search with 50 percent probability. In the benchmark assessment, TSA is compared with zebra optimization algorithm (ZOA), walrus optimization algorithm (WaOA), migration algorithm (MA), total interaction algorithm (TIA), and one-toone based optimization (OOBO). The result indicates the dominance of TSA among its comparators. TSA is better than ZOA, WaOA, MA, TIA, and OOBO in 21, 19, 17, 19, and 22 functions consecutively out of 23 functions. [ABSTRACT FROM AUTHOR]

A new swarm-based metaheuristic, namely the enriched coati osprey algorithm (ECOA), is proposed in this paper. As its name suggests, ECOA hybridizes two new metaheuristics, the coati optimization algorithm (COA) and the osprey optimization algorithm (OOA). ECOA is constructed by five searches performed sequentially by the swarm members. The first three are directed searches, while the last two are neighborhood searches. All three directed searches are adopted from COA and OOA. Meanwhile, the four-bordered neighborhood search is developed based on a new approach. During the assessment, ECOA was challenged to overcome the set of 23 functions and contended with five new metaheuristics: total interaction algorithm (TIA), golden search optimization (GSO), average and subtraction-based optimization (ASBO), COA, and OOA. The result shows that ECOA outperforms TIA, GSO, ASBO, COA, and OOA in 16, 23, 18, 21, and 21 functions. Meanwhile, the individual search test result shows that the directed searches perform better than the neighborhood searches. Moreover, the directed search toward the best member becomes the most dominant search. [ABSTRACT FROM AUTHOR]

This study designs a new stochastic optimization i.e., metaheuristic technique, namely run-catch optimizer (RCO). RCO provides a distinct mechanism regarding the diversification-intensification strategy. Each member runs two sequential activities in every iteration. The first activity is running and the second one is catching. Each activity generates a seed. In the first activity, a virtual best member moves away from the corresponding member to become the first seed. The second seed is generated along the way between the corresponding member and the first seed in the first activity. If both seeds fail to improve, the member conducts a random search to find a new member. Otherwise, the better seed replaces the corresponding member. Then, RCO is challenged to handle both theoretical and real-world optimization problems. The classic 23 functions represent theoretical problems, while the outsourcing optimization problem represents the practical problem. In these simulations, RCO is confronted with five other algorithms: grey wolf optimizer (GWO), particle swarm optimization (PSO), marine predator algorithm (MPA), Komodo mlipir algorithm (KMA), and pelican optimization algorithm (POA). The result shows that RCO is better than POA, KMA, MPA, GWO and PSO in optimizing 20, 21, 13, 12, and 21 functions consecutively. Meanwhile, RCO is better than PSO, GWO, and KMA, but worse than MPA and POA in optimizing the outsourcing problem. [ABSTRACT FROM AUTHOR]

This paper demonstrates a novel simple metaheuristic algorithm, the cone search (CS). This name comes from the distinct strategy of CS in searching the suboptimal solution. In the early iteration, its local search space is wide to facilitate exploration. The local search space is reduced linearly during the iteration so that the exploration changes to exploitation gradually. As a swarm intelligence, CS contains several autonomous agents and a collective intelligence called memory. This memory consists of several best solutions. In this work, CS is challenged to find the global optimal of 23 benchmark functions. In the simulation, CS is compared with four metaheuristic algorithms: particle swarm optimization (PSO), marine-predators algorithm (MPA), Komodo mlipir algorithm (KMA), and pelican optimization algorithm (POA). The result shows that CS performs well in solving these 23 functions. Moreover, it can find the global optimal of six fixeddimension multimodal functions: Branin, six hump camel, Hartman 6, Shekel 5, Shekel 7, and Goldstein price. CS beats PSO, MPA, KMA, and POA in solving 23, 22, 21, 22 functions respectively. [ABSTRACT FROM AUTHOR]

Job Shop Scheduling (JSS) is one of the problems in the production process. The sequence of operation and processing time are often different because some jobs consist of multiple processes, with each being performed by a different machine. The purpose of JSS is to determine the order in which jobs are processed to meet specific optimization criteria. Several solutions to Job Shop Scheduling Problem (JSSP) have been proposed, either with exact approaches or heuristics. It was observed that one of the most widely used heuristic approaches is metaheuristics, specifically the Genetic Algorithm, which has the advantage of finding solutions globally, also known as global optimal. However, this algorithm is often trapped in a search that only involves local optimal values. This study proposes an approach to improve the Genetic Algorithm performance by combining it with another metaheuristic, known as the Firefly Algorithm, which has the advantage of finding solutions locally or called local optimal. It is, therefore, possible to maintain global and local optimal balance and obtain better performance by combining these two algorithms as well. Furthermore, two approaches are proposed, which include S-GAFA and C-GAFA. Several experiments were performed to measure these proposed algorithms. It was observed that S-GAFA and C-GAFA performed better than the Genetic Algorithm in solving JSSP. [ABSTRACT FROM AUTHOR]