In the first phase of the experiment the test problems were scheduled using DPSM5 (the 5 included in the name of DPSM connotes that in this version of DPSM the number of priorities used to check the objective function for the clock cycle is five).
In order to explore the capability and also suitability of DPSM in handling the real world projects, the test projects were screened to separate the more complicated and real world projects from the theoretical and simple projects.
D1: DPSM option I using a clock cycle equal to 10 per cent of original critical path length, e.
D2: DPSM option 2 using a clock cycle equal to 10 per cent of original critical path length, e.
D3: DPSM option 3 using a clock cycle equal to 10 per cent of original critical path length e.
D4: DPSM option 4 using a clock cycle equal to 10 per cent of original critical path length, e.
The last line of data marked as DOP (DPSM Optimal results) indicates the average of the measures based on the best results produced by DPSM through a search technique.
The analysis of measures of performance for application of DPSM to the real world projects shows that generally DPSM is more efficient in tackling the practical projects.
The comparison of performance of the traditional priorities with DPSM case of clock cycle equal to 10 per cent of critical path length showed that DPSM could produce good quality reliable solutions, though it did not perform very well compared with P3.
The result was drastically changed and DPSM options almost totally outperformed the rest of priorities.
Comparing the results of scheduling the same 100 projects by two versions of DPSM, that is DPSM5 and DPSM10, showed that the performance of DPSM5 is good enough and increasing the number of priorities used in DPSM is not recommended.
The second important question tackled in this part was how DPSM deals with practical, real world complex projects.