Cuckoo Search (CS) is a meta-heuristic algorithm inspired from a breeding behavior of cuckoo. CS is one of the efficient optimization algorithm with strong search capability. CS is different from such other modern meta-heuristic algorithms in terms of incorporating L´evy flight as the search strategy. For the strong and reliable search capability, many applications of CS have been proposed in a wide range of optimization problems. Most of these applications are complex but static problems including multi-modal functions.
To tackle dynamic problems in which the optimum changes, We proposed Dynamic Cuckoo Search, which is introduce the following three additional mechanisms into CS: (a)Local Solution Comparison strategy, which keeps a diversity of the solutions; (b)Short-range Searching strategy, which works as the local search; and (c)Concurrent Solution Generating strategy, which promotes global search.
D-CS has both the solution finding and tracking capabilities in addition to the solution holding capability; and D-CS can find the optima quickly by holding the good potential solutions in all three different types of the dynamic environments.
Air traffic management: Landing route and order optimization
For safety and economic aircraft landing, it is required to optimize both the landing routes of multiple aircrafts and their landing sequence in real-time as the air transport service. Such landing route and landing sequence are important issue in the ﬁeld of air traﬃc control science because air traﬃc controllers should determine both the landing routes and their landing sequence. This problem is called as the Aircraft Landing Problem(ALP).
For this issue, the meta-heuristic methods such as Genetic Algorithm (GA) have a potential of solving diﬃcult problems. From this potential, the optimization methods based on GA were proposed. In these methods, the solutions (e.g., the landing routes of an aircraft) are converged into the best one by the evolutionary computation mechanism (i.e., the solutions are evaluated according to the ﬁtness and the solutions having the high ﬁtness value are kept in the next generation). However, such an evolutionary mechanism reduces the diversity of solutions, which is a critical problem in ALP because the diversity of solutions is indispensable to adjust the appropriate interval between the leading and following aircrafts especially in bad weather condition(note that an aircraft cannot drastically change its landing route if the evolutionary computation ﬁnds the only one best route or very similar routes).
To tackle this issue, We proposes the optimization method that optimizes the solutions (from the viewpoint of the route distance) while promoting to increase the diversity of the solutions(from the view point of a variety of the landing routes). Speciﬁcally, our proposed method employs the two-objective evaluation method based on NSGA-II to consider both optimality and diversity criteria. For the diversity issue, this paper employs the novelty metric proposed by Lehman.
Evolutionary multi-objective route and fleet assignment optimisation for regular and non-regular flights
In order to generate the flight network for a different demand of passengers in each season, airline companies consider the regular and non-regular flights separately. The regular flights are operated on the same day and time through one year, while the non-regular flights are operated on the different day and time
according to operated month. the optimisation for the regular and the non-regular
flight network is required. However, the conventional methods cannot optimise both the regular and non-regular flight networks because they are designed for only optimising the
flight network for one season (or single month) meaning that they cannot optimise it for several seasons (or multiple months).
To tackle this problem, this paper proposes the new route planning and fleet assignment method for the flight network optimisation that can assign aircrafts for both the regular and non-regular flights. For this issue, our method focuses on the priority-based GA (PriGA) with connection network model (CNM) which optimises the flight route and assignment of aircraft’s type for the regular flights by considering passenger’s demand. PriGA with CNM is based on the evolutionary computation as one of metaheuristics methods. We focus on it because the target problem is NP-hard [precisely, the FAP with more than two fleet types is NP-hard , which means that the conventional mathematical programming methods are very hard to solve the target program from a practical viewpoint (in other words, evolutionary computation is one of potential candidates for NP-hard problem to find solutions within a practical time).
For this purpose, we extend PriGA with CNM by proposing: 1) the MIN-/MAX-based fleet assignment methods that assign the aircrafts to the flight network with the lower/higher passengers’ demand in off/on peak months 2) the weight-based fleet assignment method which adjusts the number of the regular flights according to the weight values 3) the double self-adaptive fleet assignment method for both the MIN-/MAX-based and weight-based fleet assignment methods to determine an appropriate assignment (MIN or MAX) and evolve suitable weight values.