Schirmer A.: Case-Based Reasoning and Improved Adaptive Search for Project Scheduling. Naval Research Logistics, 47: 201-222, 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....study by (Hartmann and Kolisch, 2000) This study considers the following heuristics: i. three deterministic single pass heuristics with regret based random sampling from (Kolisch, 1996a,b) ii. two single pass heuristics with adaptive regret based random sampling (Kolisch and Drexl, 1996; Schirmer, 1998), iii. four genetic algorithms of (Hartmann, 1998) and (Leon and Ramamoorthy, 1995) iv. a simulated annealing algorithm of (Bouleimen and Lecocq, 1998) These heuristics were also compared with two pure random sampling methods using two di#erent heuristics for building up a schedule. The random ....

.... 600 problem instances from benchmark set j120.sm, every heuristic was allowed to construct and evaluate 5000 solutions Algorithm Reference deviation from LB in AS RCPSP 36.65 GA 1 (Hartmann, 1998) 36.74 SA (Bouleimen and Lecocq, 1998) 37.68 GA 2 (Hartmann, 1998) 38.49 adaptive sampling 1 (Schirmer, 1998) 38.70 single pass sampling 1 (Kolisch, 1996b) 38.75 single pass sampling 2 (Kolisch, 1996a,b) 38.77 adaptive sampling 2 (Kolisch and Drexl, 1996) 40.45 GA 3 (Leon and Ramamoorthy, 1995) 40.69 single pass sampling 3 (Kolisch, 1996b) 41.84 GA 4 (Hartmann, 1998) 42.25 random sampling 1 ....

Schirmer, A. (1998). Case-based reasoning and improved adaptive search for project scheduling. Naval Research Logistics. submitted.

....set has a selection probability greater than zero and thus every schedule of the population can be generated. Schirmer and Riesenberg [33] propose a variant of RBRS where ffl is determined dynamically. So called adaptive RBRS have been proposed by Kolisch and Drexl [18] as well as Schirmer [32]. The essence of adaptive sampling is to select the SGS, the priority rule, and the way the selection probabilities are calculated based on characteristics of the problem instance at hand. We refer to these characteristics, e.g. the number of activities, the resource strength, and the resource ....

....SGS with the LFT priority rule and the parallel SGS with the WCS priority rule while employing deterministic and regret based sampling activity selection. The decision on the specific method is based on an analysis of the problem at hand and the number of iterations already performed. Schirmer [32] has extended this approach by employing both schedule generation schemes together with four different priority rules and two different sampling schemes (RBRS and a variant of RBRS) 2.3 Metaheuristic Approaches Many metaheuristic strategies such as genetic algorithms, simulated annealing, and ....

[Article contains additional citation context not shown here]

A. Schirmer. Case--based reasoning and improved adaptive search for project scheduling. Technical Report 472, Manuskripte aus den Instituten fur Betriebswirtschaftslehre der Universit at Kiel, 1998.

.... [20] The tested priority rule based sampling methods include the adaptive procedure of Kolisch and Drexl [15] the latest finish time (LFT) rule and worst case slack (WCS) rule based methods of Kolisch [14] the random sampling heuristic of Kolisch [12] and the adaptive approach of Schirmer [29]. The LFT based as well as the random sampling method was tested separately with the serial and the parallel SGS. Table 1 gives the average percentage deviations from the optimal makespan for the ProGen instance set with 30 activities in a project obtained from the evaluation of 1000 and 5000 ....

....found in Table 3. In each table, the heuristics are sorted according to descending performance with respect to 5000 Iterations Algorithm reference 1000 5000 self adapting GA (new) 0.36 0.17 simulated annealing Bouleimen, Lecocq [3] 0.38 0.23 GA Hartmann [9] 0.54 0. 25 adaptive sampling Schirmer [29] 0.65 0.44 tabu search Baar et al. 1] 0.86 0.44 adaptive sampling Kolisch, Drexl [15] 0.74 0.52 serial sampling (LFT) Kolisch [14] 0.83 0.53 serial random sampling Kolisch [12] 1.44 1.00 parallel sampling (WCS) Kolisch [13, 14] 1.40 1.28 parallel sampling (LFT) Kolisch [14] 1.40 1.29 parallel ....

[Article contains additional citation context not shown here]

A. Schirmer. Case-based reasoning and improved adaptive search for project scheduling. Manuskripte aus den Instituten fur Betriebswirtschaftslehre 472, Universitat Kiel, Germany, 1998.

....reaches parameter subspaces corresponding to good regions of the solution space. Indeed, this advocates storing such parameter subspaces and using them as starting regions when tackling planning instances of future years (such an approach may be implemented as a case based reasoning system, cf. Schirmer 1998). In column four, the service level is reported. These numbers are of high significance for the operational planning of LTT since a high service level is an important requirement for achieving and maintaining market share. With regard to this objective, rules LFSP and AIPT produce the best 21 ....

SCHIRMER, A. (1998), "Case-based reasoning and improved adaptive search for project scheduling", Manuskripte aus den Instituten fr Betriebswirtschaftslehre der Universitt Kiel 472, ftp://www.wiso.uni-kiel.de/pub/operations-research/wp472.ps.

....changes in the rules employed, thus lessening the importance of expertise on the characteristics of such rules for the problem at hand. Expertise on both areas would otherwise be necessary to identify good algorithms, and usually extensive experimentation is required to acquire it (Schirmer 1998; Schirmer 2000). Third, they are simple and easy to implement. Therefore, adaptive control schemes provide a commendable approach in situations where little knowledge is available on what constitutes a good rule or what makes an instance easy or difficult to solve. The remainder of this work is structured as ....

....random sampling scheme (RBRS; cf. Drexl 1991; Drexl, Grnewald 1993) In an extensive study, Kolisch (1996) found the scheme to dominate all other randomization schemes proposed so far. We use a recent modification of this scheme (MRBRS) that outperforms even the RBRS (Schirmer 1999, pp. 52 53; Schirmer 2000). The regret value of a candidate d measures the worst case consequence that could possibly result from selecting another candidate. Let denote V(D n ) the set of priority values of all candidates in a decision set D n . Then, the regrets are computed as v (d) max ( min ( V v ....

[Article contains additional citation context not shown here]

SCHIRMER, A. (2000), "Case-based reasoning and improved adaptive search for project scheduling", to appear in Naval Research Logistics.

No context found.

Schirmer A.: Case-Based Reasoning and Improved Adaptive Search for Project Scheduling. Naval Research Logistics, 47: 201-222, 2000.

No context found.

Schirmer, A. (2000). Case-Based Reasoning and Improved Adaptive Search for Project Scheduling, Naval Research Logistics, 47, 201-222

No context found.

Schirmer A.: Case-Based Reasoning and Improved Adaptive Search for Project Scheduling. Naval Research Logistics, 47: 201-222, 2000.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC