Do: Try to talk to your party and make sure everyone’s on the same page
The party needs to be on reasonably even footing, or else it’s impossible to create one encounter that challenges them all evenly. If this is not the case, you need to discuss with the party what can be done about it.
The optimizing player can be a recurso here, provided he’s mature and helpful. Ask him to offer his help to other players if they want it. If they don’t, you need to have a discussion about what power/optimization level is desired for the game. A compromise should be possible. This may involve the psychic warrior agreeing not to use certain tricks, or the rest of the group agreeing to some retcons to their characters to improve their optimization. Most likely (this is a compromise), both.
This is an entirely out-of-character issue, and no one is wrong. The error is entirely one of a failure to communicate: people had different ideas of what is expected of their characters. By having this discussion and making a compromise, and allowing some retcons to characters, you can fix that. Particularly if the optimization-minded player is honest, mature, and interested in helping: if he knows what he’s doing, he should both know how to improve the other characters e de-optimize his own charcter.
Do: Consider houserules to improve the weaker classes that have been chosen
Cleric and wizard: should be fine
The cleric and wizard can, if they want to, wake up tomorrow, prepare better spells, and surge far ahead of the psychic warrior. There just are not options for the psychic warrior that can keep up with what those classes can do, with the right spells. This will also become more apparent as they get more levels/more powerful spells and the psychic warrior falls further behind in terms of highest-power level and total power points. So those two should be fine; they can catch up with the psychic warrior whenever they want, and will start to do so naturally.
The fighter, ranger, and rogues, however, have a lot of problems.
Fighter: bonus feats just aren’t valuable enough; go warblade if possible
For the Fighter, two levels of Fighter (i.e. a feat each level) is not a bad choice. A feat every other level, though, is just not that good, and ultimately there are only so many fighter feats you should really need, which means after Fighter 2 the class should be about done. So ultimately, you don’t necessarily need to houserule the fighter, just... recognize that it’s not a class to take until 20.
Thus, I would strongly recommend retconning the third Fighter level, and multiclassing to something else. A level of barbarian does wonders, particularly with Lion Spirit Totem from Campeão Completo, as does the warblade class from Tomo da Batalha.
Alternativamente, Dungeonscape has a really solid dungeoncrasher fighter variant, which does tons of damage when bull rushing someone into a wall, so that extends the fighter class out to 6th level, and the Zhentarim soldier variant from the Campeões da bravura aprimoramento da web makes Fighter 9 an excellent level for an Intimidation build.
But mostly, the warblade from Tomo da Batalha is everything the fighter should have been, but done right, and I really recommend that your player consider it. It will do a lot to help. The best part about the warblade is that you can get it completamente grátis, along with the full text of all the maneuvers. All you need to know is that warblade Initiator Level is equal to warblade level + half your other levels, and you need the same Initiator Level to pick a maneuver as a wizard would need Caster Level to pick a spell.
Ranger: needs damage, better pet, better spells
The ranger class... just doesn’t have a lot going for it. It has full-BAB, good HD, and lots of skills, which is agradável, but lackluster class features mess that up. Track is kind of gimmicky; whether you track or not is usually more a matter of plot than feats. Favored Enemy bonuses are extremamente small and conditional, so most people forget they have them. Ranger spells are decent enough, but too little, too late. The bonus Combat Style feats allow lots of attacks, which favor combatants who have lots of bonus damage, but rangers don’t. And the Animal Companion is so low-level that it can’t do more than scout.
A good solution to the ranger’s damage problems are to use the Swift Hunter feat from Canalha completo along with four levels in Aventureiro Completo’s scout class. Another thing I like to do is swap the ranger and druid’s Animal Companions, so the ranger gets a full, big pet, while the powerful spellcasting druid gets the scout. Finally, giving the ranger (and paladin) a spell progression like the bard’s (you’ll have to make up some 5th and 6th level spell lists) does a lot to help.
Rogues: Too many things immune to damage, not enough mobility
The rogues aren’t as bad off as the previous two, but the number of things immune to Sneak Attack are a huge problem. I’d suggest eliminating the automatic immunity of Constructs, Elementals, Plants, and Undead, as a start. Oozes aren’t super-common and tend to be kind of “puzzle monsters” anyway, so those can probably stay immune.
Also, it would be really nice if the rogue could get a way to move, and attack repeatedly. Maybe a rogue special ability that works like Spring Attack, except that you get all your attacks? Might be too much, but then there are ways to get that anyway (and the psychic warrior definitely can move and attack with encontrão or carga do leão psiônico).
Don’t: Engineer situations where the psychic warrior is useless
Right now you have a problem where the psychic warrior outshines everyone else. You poderia engineer situations that require the other classes’ specific talents and he can’t do anything. This solves nothing. It’s basically “giving up” on the idea that everyone can play together at the same time and contribute meaningfully; it feels forced and gimmicky (because it is forced and gimmicky). Furthermore, “two wrongs don’t make a right” – having todos take a turn being useless isn’t really better than someone managing to avoid that. It’s a game, no one wants to be told “oh, you don’t have Trapfinding, so you don’t get to play for this next bit.” It’s a non-solution that I cannot recommend more strongly against.