Common use of Implementation & Performance Clause in Contracts

Implementation & Performance. The optimisation is implemented in the HiP-HOPS tool, building on its existing support for optimisation and the prototype ASIL decomposition algorithm described earlier. As a result, performance between the standard exhaustive algorithm and the optimisation approaches can be directly compared. The central advantage of the optimisation approach is that it does not require HiP-HOPS to consider every possible solution, meaning that it performs much, much quicker; the trade-off is that it will not discover every solution in each case. The aim is simply to obtain a reasonable number of good, valid solutions. For the small six component example described earlier, both the exhaustive algorithm and the optimisation algorithms are very quick (<1 second), and in the case of the optimisation, it also finds the valid solution. For larger examples, the exhaustive algorithm performs much worse, but will usually find all valid solutions eventually (unless the algorithm runs out of memory or is cancelled early). The optimisation algorithms scale far better but are not guaranteed to find any valid solutions, although the penalty-based approach has thus far been more successful. In one medium-sized example, the penalty-based optimisation took 18 seconds versus the exhaustive approach’s 18 hours, but the exhaustive algorithm did find all 50 valid solutions whereas the optimisation found only one. Continued testing and development is required to try to improve the success rate of the optimisation algorithms. Their advantage in performance is clear as optimisation is far more scalable than standard exhaustive approaches; however, to be more practical, they need tweaking further to yield more valid solutions. In addition, further developments of the standard exhaustive algorithm may also produce good results, although it is doubtful that it will ever be fast enough to be applied to large systems.