Evolutionary Non-linear Great Deluge for University Course Timetabling

Abstract

Abstract. This paper presents a hybrid evolutionary algorithm to tackle university course timetabling problems. The proposed approach is an extension of a non-linear great deluge algorithm in which evolutionary operators are incorporated. First, we generate a population of feasible solutions using a tailored process that incorporates heuristics for graph colouring and assignment problems. That initialisation process is capable of producing feasible solutions even for the large and most constrained problem instances. Then, the population of feasible timetables is subject to a steady-state evolutionary process that combines mutation and stochastic local search. We conduct experiments to evaluate the performance of the proposed hybrid algorithm and in particular, the contribution of the evolutionary operators. Our results show that the hybrid between non-linear great deluge and evolutionary operators produces very good results on the instances of the university course timetabling problem tackled here.

Hybrid of genetic algorithm and great deluge algorithm for rough set attribute reduction

“…The GD algorithm controls the search space using a boundary 'level' and it accepts 2 kinds of solutions: the 1st is the best solution, when there is an improvement in the quality, and the 2nd is a worse solution, when the quality has a better standing compared to the current level. The GD algorithm has shown its performance in solving many optimization problems (e.g., [15][16][17]). The GD algorithm for rough set attribute reduction (GD-RSAR) was presented in 2010 [18].…”

A great deluge and tabu search hybrid with two-stage memory support for quadratic assignment problem

“…Scores for taixxa instances are available for all metaheuristics, except GRASP, and TMSGD-QAP performs equally well or better than its competitors in 4 of these Table 9 Comparisons with literature for type-I QAPLIB benchmark problems. 0.000 (10) 0.000 0.000 (10) 0.000 0.000 0.110 0.000 0.061 0.000 0.191 tai25a 0.000 (10) 0.000 0.000 (10) 0.000 0.000 0.290 0.037 0.088 0.000 0.000 0.488 tai30a 0.091 (6) 0.000 0.000 (10) 0.000 0.000 0.340 0.003 0.019 0.000 0.000 0.359 tai35a 0.153 (2) 0.000 0.000 (10) 9 QAP instances, whereas ETS1 and GA/TS/I are the two best performing algorithms for the remaining taixxa instances. While the proposed algorithm is better than RTS for 5 of 9 taixxa problems, it outperforms ACO2 for all problem instances.…”

“…The author claimed that the proposed combination generated consistently good results for benchmark problems used in experimental studies. Landa-Silva et al proposed another hybrid of evolutionary algorithms and NLGDA where an individual selected from the current population by tournament selection is modified by mutation and the mutated individual is improved by NLGDA [10]. If the resulting solution is better than the worst solution in GA's population, it replaces that worst solution.…”

“…Table 14 illustrates the APD scores for the final set of QAP instances containing the difficult drexx problems. Unfortunately, only one algorithm is found in literature with scores of some of (8) 0.175 (14) 5.915 10.100 Tai75e02 4.165 (8) 0.339 (14) 8.751 Tai75e03 1.081 (14) 0.000 (20) 2.681 Tai75e04 1.005 (16) 0.030 (19) 3.705 Tai75e05 1.412 (11) 0.000 (20) 2.412 Tai75e06 2.115 (10) 0.000 (20) 4.005 Tai75e07 4.662 (7) 0.136 (17) 8.015 Tai75e08…”

Designing a multi-agent approach system for distributed course timetabling

“…Kapalsky et al [6] tackled a real-world Distributed Timetabling Problem (DisTTP) using a multi-agent system paradigm. Each agent in their model has a different set of requirements to guide them in their search for the optimal solution.…”