Research Article

Enhancing Hyperheuristics for the Knapsack Problem through Fuzzy Logic

Table 3

Results from the hyperheuristics with the 2400 hard instances.

MethodsHard Pisinger instances
2050100200

Default380427195629591898543937922860
MaxP372458892698411838293036723430
MinW3867345100551482028761640898138
MaxPW4039708103535552079162741776413
Oracle4084003103808512080228141783753
HH4Average3949205.7610048476.7020107840.3340333536.33
Standard deviation99734.35 (2.52%)380935.64 (3.79%)874132.26 (4.34%)1871087.96 (4.63%)
Best4045507103556582079162741776413
Worst373457193128901850216536963737

HH6Average3950021.5010079568.6320193352.1040525684.03
Standard deviation84499.05 (2.13%)324209.91 (3.21%)747259.32 (3.70%)1601814.81 (3.95%)
Best4038167103500202078676841775010
Worst374161793141301850204136957315

HH8Average3957540.7610089769.2320192245.8640509777.40
Standard deviation92317.06 (2.33%)333676.40 (3.30%)777075.35 (3.84%)1681688.34 (4.15%)
Best4042198103537052079162741776413
Worst374197893075301830301636466476

Binary GAAverage4073227.769755928.9018475255.4335069085.23
Standard deviation3096.26 (0.07%)7888.44 (0.08%)17760.93 (0.09%)41037.09 (0.11%)
Best407872397678181850354535167446
Worst406796897380181843389534989419

FuzzyHHAverage4008457.7610175892.1620342907.4040794413.63
Standard deviation43384.17 (1.08%)155492.79 (1.52%)359855.26 (1.76%)772999.81 (1.89%)
Best4050373103556102077793341757401
Worst388622597870821947716339003392

Results in bold are the best.