Production fine planning using a solution archive of priority rules

Abstract

Production Fine Planning is often performed directly using all information and assuming that it is fixed. In practice, however, this information changes regularly and the plan has to be adapted. This often means a complete rescheduling of all operations. We present a new approach to this problem by optimizing priority rules that can sort the available next actions. These priority rules often yield similar results even though they do not resemble each other. By using genetic programming to build these priority rules, a distributed system to compute the simulations and a solution archive with a cache of hundreds of thousands of priority rules, new insights into priority rule-based optimization are gained. This archive does not only speed up calculation by avoiding re-simulation of the same rule but can provide a pseudo Pareto front of shorter sub-optimal solutions that facilitate interpretation of the more complex rules and their evolution during the optimization process.

Automated Design of Production Scheduling Heuristics: A Review

“…Shop configuration Single machine [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33] Parallel machines [34] Job shop [26], [31], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56] Flexible job shop [57], [58], [59], [60], [61], [62], [63], [64], [65] Flow shop [25], [66], [67] Special processing Sequence-dependent setups …”

“…Parametric representation [21], [38], [42], [46] [35], [36], [41], [45], [47], [52], [57], [58], [59], [67] Grammar-based representation [24], [30], [39] [22], [23], [25], [26], [27], [28], [29], [31], [32], [33], [34], [37], [40], [43], [44], [48], [49], [50], [51], [52], [53], [54], [55], [56], [60], [61], [62], [63], [64], [65], [66] The next section (III-A) discusses the two types of learning methods used within hyper-heuristics, followed by a discussion of the selection of attributes to be provided to the hyperheuristic in Section III-B. The different representations of priority functions are presented in Section III-C together with suitable optimisation algorithms as they are closely tied to the chosen representation.…”

“…Table IV lists the functions that have been used within hyper-heuristics. It shows that the four basic arithmetic operators are included in the function set in every of the reviewed papers, with the division either implemented as protected (returns 1 if divisor is 0) or unprotected (returns a very [23], [25], [26], [27], [29], [31], [32], [33], [34], [37], [40], [44], [43], [48], [49], [50], [51], [52], [53], [54], [55], [56], [60], [61], [62], [63], [64], [65], [66] Logical ifte (ternary) [32], [33], [44], [43], [49], [50], [51], [52], [53], [54], [55], [56], [61], [63], [65] ifte (quaternary) [25], [26], …”

Architecture and Design of the HeuristicLab Optimization Environment

“…For example, the simulation of a production facility might require to solve a scheduling problem to determine the best order of job executions which in turn requires the integration of an optimization approach. HeuristicLab has been used in both cases [32,39] successfully, but while the first case has been generalized and abstracted as shown in this work, the second case still requires a tighter coupling with the simulation model. The generalization and abstraction of the second case is a topic for future work.…”

“…To further speed up the optimization the user can add aforementioned EvaluationCache and the respective evaluator. The cache can be persisted to a file or exported as a comma-separated-values (CSV) file for later analysis [32].…”