Citation Context

...the experiments in this paper we define the parameters ϕt and δt as: and ϕt(hj) = { 0.99, if an improving move is made max {ϕt−1 − 0.01, 0.01}, if a non-improving move is made (6) δt(hj) = 1 − ϕt(hj) =-=(7)-=- 4 Computational Results Prior to the original competition, the results of eight hyper-heuristics were provided by the organisers to assess an algorithms performance [18]. These hyperheuristics were i...

Citation Context

...of real-word optimisation problems such as bin packing [7], [8], dynamic environments [9], [10], knapsack problems [11], scheduling [4], [5], [7], [12], [13], timetabling [8], [12], [14], [15], [16], =-=[17]-=-, [18], [19] and vehicle routing [7], [20], [21]. The HyFlex [22] framework was initially developed for the first Cross-domain Heuristic Search Challenge (CHeSC) [7] and is a software framework “desig...

Citation Context

...een examinations. iii. As much as possible, students should not be scheduled to sit more than one examination in a day 3. The Algorithm The algorithm constitutes two phases as previously explained. In the first phase we construct an initial feasible solution. This feasible solution is gradually improved in the second phase. 3.1 Phase I: Seek an Initial Solution In this phase we employ a sequential graph colouring heuristic to construct an initial solution. The relationship between the examination timetabling and graph colouring problems is widely discussed in the literature (See [8], [12] and [13]). Basically, the examination timetabling problem is represented as a graph colouring problem where the vertices represent examinations, the edges represent conflicts between two examinations and colour of the vertices represent different timeslots in the timetable. In order to construct a conflict-free timetable using graph based algorithm, a list of examinations is generated according to the “difficulty” in scheduling each examination which depends on certain criteria. Some common graph based heuristics are Largest Enrolment First, Largest Degree First, Saturation Degree First, and Largest C...

Citation Context

...ate heuristics for timetabling. Specifically, the system tries to discover a set of labelled points in the space of the problem states. Each label refers to a heuristic, and the algorithm works by repeatedly finding the nearest labelled point to the current condition and applies its label until a complete solution is built. Various different forms of problem-state description and methods of measuring the fitness were studied. The approach was able to generate fast and simple problem-solving algorithms that offer good performance over a range of exam and class timetabling problems. Sabar et al [43] utilise hierarchical hybridisations of four low-level graph colouring heuristics for producing even orderings. A combined difficulty index is calculated by considering all the orderings and events are scheduled according to this index. The approach produced competitive result on the studied benchmark instances. Numbers of authors have done work in the area of scheduling algorithms. Table 1 represents the comparative study of various scheduling algorithm, nature of scheduling algorithm, objective criteria i.e. the parameters which have been focused for optimization and the environment in which...